{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "a4dfab9b-ba00-4c13-92ec-cd4c3eb68eec",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt \n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "517c5f3f-b915-46e4-a9d4-4476d0a5e3c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_X_raw = pd.read_excel('masbasis_spectral_complete.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "c9726185-5dba-4bb2-b17d-d0c72d9dc5aa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>location</th>\n",
       "      <th>plot_number</th>\n",
       "      <th>plot_type</th>\n",
       "      <th>camera</th>\n",
       "      <th>line_number</th>\n",
       "      <th>masbasis2015</th>\n",
       "      <th>line</th>\n",
       "      <th>rep</th>\n",
       "      <th>block</th>\n",
       "      <th>column</th>\n",
       "      <th>...</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "      <th>blue_mean</th>\n",
       "      <th>green_mean</th>\n",
       "      <th>red_mean</th>\n",
       "      <th>rededge_mean</th>\n",
       "      <th>nir_mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>10000</th>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1162.0</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>GN14515</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.016666</td>\n",
       "      <td>0.041038</td>\n",
       "      <td>0.021821</td>\n",
       "      <td>0.107767</td>\n",
       "      <td>0.417709</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10001</th>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1162.0</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>GN14515</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.003375</td>\n",
       "      <td>0.007734</td>\n",
       "      <td>0.004212</td>\n",
       "      <td>0.019883</td>\n",
       "      <td>0.078168</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10002</th>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1162.0</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>GN14515</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.019329</td>\n",
       "      <td>0.050943</td>\n",
       "      <td>0.029267</td>\n",
       "      <td>0.112232</td>\n",
       "      <td>0.349410</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10003</th>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1162.0</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>GN14515</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.023749</td>\n",
       "      <td>0.066180</td>\n",
       "      <td>0.042014</td>\n",
       "      <td>0.141552</td>\n",
       "      <td>0.347312</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10004</th>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1162.0</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>1624.0</td>\n",
       "      <td>GN14515</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.026614</td>\n",
       "      <td>0.070786</td>\n",
       "      <td>0.067438</td>\n",
       "      <td>0.166831</td>\n",
       "      <td>0.296682</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        location  plot_number plot_type camera  line_number  masbasis2015  \\\n",
       "10000  vollebekk       1162.0     yield   mica       1624.0        1624.0   \n",
       "10001  vollebekk       1162.0     yield   mica       1624.0        1624.0   \n",
       "10002  vollebekk       1162.0     yield   mica       1624.0        1624.0   \n",
       "10003  vollebekk       1162.0     yield   mica       1624.0        1624.0   \n",
       "10004  vollebekk       1162.0     yield   mica       1624.0        1624.0   \n",
       "\n",
       "          line  rep  block  column  ...  blue_median green_median  red_median  \\\n",
       "10000  GN14515  1.0    9.0    62.0  ...     0.016666     0.041038    0.021821   \n",
       "10001  GN14515  1.0    9.0    62.0  ...     0.003375     0.007734    0.004212   \n",
       "10002  GN14515  1.0    9.0    62.0  ...     0.019329     0.050943    0.029267   \n",
       "10003  GN14515  1.0    9.0    62.0  ...     0.023749     0.066180    0.042014   \n",
       "10004  GN14515  1.0    9.0    62.0  ...     0.026614     0.070786    0.067438   \n",
       "\n",
       "       rededge_median  nir_median  blue_mean  green_mean  red_mean  \\\n",
       "10000        0.107767    0.417709        NaN         NaN       NaN   \n",
       "10001        0.019883    0.078168        NaN         NaN       NaN   \n",
       "10002        0.112232    0.349410        NaN         NaN       NaN   \n",
       "10003        0.141552    0.347312        NaN         NaN       NaN   \n",
       "10004        0.166831    0.296682        NaN         NaN       NaN   \n",
       "\n",
       "       rededge_mean  nir_mean  \n",
       "10000           NaN       NaN  \n",
       "10001           NaN       NaN  \n",
       "10002           NaN       NaN  \n",
       "10003           NaN       NaN  \n",
       "10004           NaN       NaN  \n",
       "\n",
       "[5 rows x 22 columns]"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X_raw['date'] = pd.to_datetime(df_X_raw['date'])\n",
    "df_X_raw.iloc[10000:10005,2:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "0f594b2e-ec11-4446-8fa9-f868246f3128",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(57264, 24)"
      ]
     },
     "execution_count": 148,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X_raw.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "d88003f5-67a5-4ecc-865f-09dc21e4fad9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(57264, 13)"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X1 = df_X_raw[['env','season','location','plot_number','plot_type','camera','lodging','date','blue_median','green_median','red_median','rededge_median','nir_median']]\n",
    "df_X1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "8dfa0161-ebf6-49db-90dd-2824c04da0d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_ = df_X1.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "7974abbd-2bbc-4290-9f0d-225a5981e703",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_['date'] = pd.to_datetime(df_['date'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "cc8ccf99-ee4c-4f9b-8012-214d7285917d",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Remove Lodged Plots\n",
    "df_X = df_.loc[(df_X1['lodging'] == False) & (df_X1['plot_type'] == 'yield') & (df_X1['location'] == 'vollebekk') ]\n",
    "df_X.shape\n",
    "df_X.set_index('plot_number', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "id": "8586b509-609b-4ec1-962a-b3fbee1758a0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "576"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X.isna().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "0c782bb8-d6b8-4272-8d54-dae933beb5ca",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3080, 12)"
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X_2019 = df_X.loc[df_X['season'] == 2019]\n",
    "\n",
    "df_X_2019.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "4440ed2d-117b-4f44-b8a5-d598042fc5b9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6048, 12)"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X_2020 = df_X.loc[(df_X['season'] == 2020)]\n",
    "\n",
    "df_X_2020.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "id": "735c1e14-3e12-4767-9067-b9daa523ee8d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>env</th>\n",
       "      <th>season</th>\n",
       "      <th>location</th>\n",
       "      <th>plot_type</th>\n",
       "      <th>camera</th>\n",
       "      <th>lodging</th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>False</td>\n",
       "      <td>2020-06-18</td>\n",
       "      <td>0.015137</td>\n",
       "      <td>0.037526</td>\n",
       "      <td>0.014911</td>\n",
       "      <td>0.109639</td>\n",
       "      <td>0.431767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>False</td>\n",
       "      <td>2020-06-24</td>\n",
       "      <td>0.014972</td>\n",
       "      <td>0.034236</td>\n",
       "      <td>0.016068</td>\n",
       "      <td>0.094503</td>\n",
       "      <td>0.391461</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>False</td>\n",
       "      <td>2020-06-26</td>\n",
       "      <td>0.012624</td>\n",
       "      <td>0.027086</td>\n",
       "      <td>0.015051</td>\n",
       "      <td>0.075851</td>\n",
       "      <td>0.435237</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>False</td>\n",
       "      <td>2020-07-01</td>\n",
       "      <td>0.017417</td>\n",
       "      <td>0.043442</td>\n",
       "      <td>0.020844</td>\n",
       "      <td>0.107552</td>\n",
       "      <td>0.433779</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>yield</td>\n",
       "      <td>mica</td>\n",
       "      <td>False</td>\n",
       "      <td>2020-07-08</td>\n",
       "      <td>0.022297</td>\n",
       "      <td>0.048837</td>\n",
       "      <td>0.020287</td>\n",
       "      <td>0.114728</td>\n",
       "      <td>0.444705</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        env  season   location plot_type camera  lodging  \\\n",
       "plot_number                                                                \n",
       "1105.0       2020_vollebekk    2020  vollebekk     yield   mica    False   \n",
       "1105.0       2020_vollebekk    2020  vollebekk     yield   mica    False   \n",
       "1105.0       2020_vollebekk    2020  vollebekk     yield   mica    False   \n",
       "1105.0       2020_vollebekk    2020  vollebekk     yield   mica    False   \n",
       "1105.0       2020_vollebekk    2020  vollebekk     yield   mica    False   \n",
       "\n",
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1105.0      2020-06-18     0.015137      0.037526    0.014911        0.109639   \n",
       "1105.0      2020-06-24     0.014972      0.034236    0.016068        0.094503   \n",
       "1105.0      2020-06-26     0.012624      0.027086    0.015051        0.075851   \n",
       "1105.0      2020-07-01     0.017417      0.043442    0.020844        0.107552   \n",
       "1105.0      2020-07-08     0.022297      0.048837    0.020287        0.114728   \n",
       "\n",
       "             nir_median  \n",
       "plot_number              \n",
       "1105.0         0.431767  \n",
       "1105.0         0.391461  \n",
       "1105.0         0.435237  \n",
       "1105.0         0.433779  \n",
       "1105.0         0.444705  "
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X_2020.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "adfcedf2-5df9-4946-81ab-b857ed8add72",
   "metadata": {},
   "source": [
    "## Data Resampling and Interpolation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "f2c6e8cf-47b0-4323-a6bc-fba9caf8c879",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.interpolate import CubicSpline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "id": "9ded61ff-9a12-4a92-941e-965575721bf1",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_20X = df_X_2020[['date','blue_median','green_median','red_median','rededge_median','nir_median']]\n",
    "df_20X.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "id": "013cf1b5-4a8f-41c1-b488-ce1330fb290a",
   "metadata": {},
   "outputs": [],
   "source": [
    "freq = '5D'\n",
    "\n",
    "df_20res = pd.DataFrame()\n",
    "\n",
    "for plot_num in df_20X['plot_number'].unique():\n",
    "    \n",
    "    data = df_20X.loc[df_20X['plot_number'] == plot_num]\n",
    "    data.set_index('date', inplace = True)\n",
    "    data_res = data.resample(freq).mean().reset_index()\n",
    "    data_res['plot_number'] = plot_num\n",
    "    \n",
    "    df_20res = pd.concat([df_20res, data_res], ignore_index=False)\n",
    "\n",
    "df_20res.set_index('plot_number', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "id": "095dbab5-a7a0-4b40-aa17-2d81aff7997a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6048, 6)"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_20res.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "id": "b9571b8e-644e-4322-867f-269300a6da53",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-06-18</td>\n",
       "      <td>0.015137</td>\n",
       "      <td>0.037526</td>\n",
       "      <td>0.014911</td>\n",
       "      <td>0.109639</td>\n",
       "      <td>0.431767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-06-23</td>\n",
       "      <td>0.013798</td>\n",
       "      <td>0.030661</td>\n",
       "      <td>0.015559</td>\n",
       "      <td>0.085177</td>\n",
       "      <td>0.413349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-06-28</td>\n",
       "      <td>0.017417</td>\n",
       "      <td>0.043442</td>\n",
       "      <td>0.020844</td>\n",
       "      <td>0.107552</td>\n",
       "      <td>0.433779</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-03</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-08</td>\n",
       "      <td>0.022297</td>\n",
       "      <td>0.048837</td>\n",
       "      <td>0.020287</td>\n",
       "      <td>0.114728</td>\n",
       "      <td>0.444705</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-13</td>\n",
       "      <td>0.011046</td>\n",
       "      <td>0.031304</td>\n",
       "      <td>0.016580</td>\n",
       "      <td>0.070497</td>\n",
       "      <td>0.232059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-18</td>\n",
       "      <td>0.022397</td>\n",
       "      <td>0.060377</td>\n",
       "      <td>0.048461</td>\n",
       "      <td>0.138072</td>\n",
       "      <td>0.290974</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-23</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-28</td>\n",
       "      <td>0.027139</td>\n",
       "      <td>0.075271</td>\n",
       "      <td>0.070734</td>\n",
       "      <td>0.173504</td>\n",
       "      <td>0.311046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-02</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-07</td>\n",
       "      <td>0.031707</td>\n",
       "      <td>0.071970</td>\n",
       "      <td>0.110121</td>\n",
       "      <td>0.155774</td>\n",
       "      <td>0.253116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-12</td>\n",
       "      <td>0.042600</td>\n",
       "      <td>0.077703</td>\n",
       "      <td>0.126265</td>\n",
       "      <td>0.151689</td>\n",
       "      <td>0.252579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>2020-06-18</td>\n",
       "      <td>0.013736</td>\n",
       "      <td>0.036581</td>\n",
       "      <td>0.013466</td>\n",
       "      <td>0.110889</td>\n",
       "      <td>0.448822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>2020-06-23</td>\n",
       "      <td>0.013585</td>\n",
       "      <td>0.031790</td>\n",
       "      <td>0.015338</td>\n",
       "      <td>0.087899</td>\n",
       "      <td>0.414929</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1105.0      2020-06-18     0.015137      0.037526    0.014911        0.109639   \n",
       "1105.0      2020-06-23     0.013798      0.030661    0.015559        0.085177   \n",
       "1105.0      2020-06-28     0.017417      0.043442    0.020844        0.107552   \n",
       "1105.0      2020-07-03          NaN           NaN         NaN             NaN   \n",
       "1105.0      2020-07-08     0.022297      0.048837    0.020287        0.114728   \n",
       "1105.0      2020-07-13     0.011046      0.031304    0.016580        0.070497   \n",
       "1105.0      2020-07-18     0.022397      0.060377    0.048461        0.138072   \n",
       "1105.0      2020-07-23          NaN           NaN         NaN             NaN   \n",
       "1105.0      2020-07-28     0.027139      0.075271    0.070734        0.173504   \n",
       "1105.0      2020-08-02          NaN           NaN         NaN             NaN   \n",
       "1105.0      2020-08-07     0.031707      0.071970    0.110121        0.155774   \n",
       "1105.0      2020-08-12     0.042600      0.077703    0.126265        0.151689   \n",
       "1106.0      2020-06-18     0.013736      0.036581    0.013466        0.110889   \n",
       "1106.0      2020-06-23     0.013585      0.031790    0.015338        0.087899   \n",
       "\n",
       "             nir_median  \n",
       "plot_number              \n",
       "1105.0         0.431767  \n",
       "1105.0         0.413349  \n",
       "1105.0         0.433779  \n",
       "1105.0              NaN  \n",
       "1105.0         0.444705  \n",
       "1105.0         0.232059  \n",
       "1105.0         0.290974  \n",
       "1105.0              NaN  \n",
       "1105.0         0.311046  \n",
       "1105.0              NaN  \n",
       "1105.0         0.253116  \n",
       "1105.0         0.252579  \n",
       "1106.0         0.448822  \n",
       "1106.0         0.414929  "
      ]
     },
     "execution_count": 166,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_20res.head(14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "id": "1b7ecaf9-cff5-43ab-b6c7-728821f70d19",
   "metadata": {},
   "outputs": [],
   "source": [
    "for column in df_20res.columns:\n",
    "    if (column != 'date'):\n",
    "        mask = df_20res[column].isnull()  # Create a mask for missing values\n",
    "        x = np.arange(len(df_20res))\n",
    "        y = df_20res[column].values\n",
    "\n",
    "        # Create a cubic spline interpolation function\n",
    "        spline = CubicSpline(x[~mask], y[~mask])\n",
    "\n",
    "        # Fill missing values with interpolated values\n",
    "        df_20res.loc[mask, column] = spline(x[mask])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "id": "5eb411e9-f0b8-48c9-9f8a-b5a029d7e40a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-06-18</td>\n",
       "      <td>0.015137</td>\n",
       "      <td>0.037526</td>\n",
       "      <td>0.014911</td>\n",
       "      <td>0.109639</td>\n",
       "      <td>0.431767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-06-23</td>\n",
       "      <td>0.013798</td>\n",
       "      <td>0.030661</td>\n",
       "      <td>0.015559</td>\n",
       "      <td>0.085177</td>\n",
       "      <td>0.413349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-06-28</td>\n",
       "      <td>0.017417</td>\n",
       "      <td>0.043442</td>\n",
       "      <td>0.020844</td>\n",
       "      <td>0.107552</td>\n",
       "      <td>0.433779</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-03</td>\n",
       "      <td>0.024159</td>\n",
       "      <td>0.055160</td>\n",
       "      <td>0.024431</td>\n",
       "      <td>0.131636</td>\n",
       "      <td>0.500632</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-08</td>\n",
       "      <td>0.022297</td>\n",
       "      <td>0.048837</td>\n",
       "      <td>0.020287</td>\n",
       "      <td>0.114728</td>\n",
       "      <td>0.444705</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-13</td>\n",
       "      <td>0.011046</td>\n",
       "      <td>0.031304</td>\n",
       "      <td>0.016580</td>\n",
       "      <td>0.070497</td>\n",
       "      <td>0.232059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-18</td>\n",
       "      <td>0.022397</td>\n",
       "      <td>0.060377</td>\n",
       "      <td>0.048461</td>\n",
       "      <td>0.138072</td>\n",
       "      <td>0.290974</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-23</td>\n",
       "      <td>0.029142</td>\n",
       "      <td>0.077585</td>\n",
       "      <td>0.066680</td>\n",
       "      <td>0.177520</td>\n",
       "      <td>0.331992</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-28</td>\n",
       "      <td>0.027139</td>\n",
       "      <td>0.075271</td>\n",
       "      <td>0.070734</td>\n",
       "      <td>0.173504</td>\n",
       "      <td>0.311046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-02</td>\n",
       "      <td>0.024164</td>\n",
       "      <td>0.068506</td>\n",
       "      <td>0.079378</td>\n",
       "      <td>0.161215</td>\n",
       "      <td>0.289961</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-07</td>\n",
       "      <td>0.031707</td>\n",
       "      <td>0.071970</td>\n",
       "      <td>0.110121</td>\n",
       "      <td>0.155774</td>\n",
       "      <td>0.253116</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-12</td>\n",
       "      <td>0.042600</td>\n",
       "      <td>0.077703</td>\n",
       "      <td>0.126265</td>\n",
       "      <td>0.151689</td>\n",
       "      <td>0.252579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>2020-06-18</td>\n",
       "      <td>0.013736</td>\n",
       "      <td>0.036581</td>\n",
       "      <td>0.013466</td>\n",
       "      <td>0.110889</td>\n",
       "      <td>0.448822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>2020-06-23</td>\n",
       "      <td>0.013585</td>\n",
       "      <td>0.031790</td>\n",
       "      <td>0.015338</td>\n",
       "      <td>0.087899</td>\n",
       "      <td>0.414929</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1105.0      2020-06-18     0.015137      0.037526    0.014911        0.109639   \n",
       "1105.0      2020-06-23     0.013798      0.030661    0.015559        0.085177   \n",
       "1105.0      2020-06-28     0.017417      0.043442    0.020844        0.107552   \n",
       "1105.0      2020-07-03     0.024159      0.055160    0.024431        0.131636   \n",
       "1105.0      2020-07-08     0.022297      0.048837    0.020287        0.114728   \n",
       "1105.0      2020-07-13     0.011046      0.031304    0.016580        0.070497   \n",
       "1105.0      2020-07-18     0.022397      0.060377    0.048461        0.138072   \n",
       "1105.0      2020-07-23     0.029142      0.077585    0.066680        0.177520   \n",
       "1105.0      2020-07-28     0.027139      0.075271    0.070734        0.173504   \n",
       "1105.0      2020-08-02     0.024164      0.068506    0.079378        0.161215   \n",
       "1105.0      2020-08-07     0.031707      0.071970    0.110121        0.155774   \n",
       "1105.0      2020-08-12     0.042600      0.077703    0.126265        0.151689   \n",
       "1106.0      2020-06-18     0.013736      0.036581    0.013466        0.110889   \n",
       "1106.0      2020-06-23     0.013585      0.031790    0.015338        0.087899   \n",
       "\n",
       "             nir_median  \n",
       "plot_number              \n",
       "1105.0         0.431767  \n",
       "1105.0         0.413349  \n",
       "1105.0         0.433779  \n",
       "1105.0         0.500632  \n",
       "1105.0         0.444705  \n",
       "1105.0         0.232059  \n",
       "1105.0         0.290974  \n",
       "1105.0         0.331992  \n",
       "1105.0         0.311046  \n",
       "1105.0         0.289961  \n",
       "1105.0         0.253116  \n",
       "1105.0         0.252579  \n",
       "1106.0         0.448822  \n",
       "1106.0         0.414929  "
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#df_20res = df_20res.fillna(df_20res.median(axis=0))\n",
    "df_20res.head(14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ede43eb6-be29-47ca-9169-419bc59d12a7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aebd901b-43a1-4aaa-a738-e763083a0daa",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "id": "e92c618e-b1dc-4d24-957a-36766402957a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from statsmodels.tsa.stattools import adfuller\n",
    "import statsmodels.api as sm\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "id": "cda6a969-83cb-40de-b37e-2fecbac42257",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "p-value for blue_median: 5.002215085193975e-12\n",
      "p-value for green_median: 2.3240879759245744e-15\n",
      "p-value for red_median: 1.3880939329816815e-10\n",
      "p-value for rededge_median: 3.4970355850164568e-16\n",
      "p-value for nir_median: 1.9332601943675917e-10\n"
     ]
    }
   ],
   "source": [
    "df20vals = df_20res.iloc[:,-5:]\n",
    "\n",
    "for col in df20vals.columns:\n",
    "    data = df20vals[col].values.reshape(-1,1)\n",
    "    result = adfuller(data)\n",
    "    \n",
    "    #print(f\"ADF Statistics for {col}: {result[0]}\")\n",
    "    print(f\"p-value for {col}: {result[1]}\")\n",
    "    #print('Critical Values:', result[4])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9c27ead7-5a4a-43aa-bff7-fd2d0bd3d681",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "id": "f617428c-9279-4f61-b9b0-89934c2464c2",
   "metadata": {},
   "outputs": [],
   "source": [
    "start_date = '2023-06-18'\n",
    "end_date = '2023-08-12'\n",
    "freq = '5D'\n",
    "date_range = pd.date_range(start=start_date, end=end_date, freq=freq)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92fc2471-d472-4888-8b22-ead2a7d0cabf",
   "metadata": {},
   "source": [
    "## Removing Outliers from each year using IQR Percentile\n",
    "Before removing the trend, we remove the outliers to possibly see an improvement in the model's accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "id": "86271d2f-1819-44c1-9f1d-31ad60f64821",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_20 = df_20res.copy()\n",
    "df_20 = df_20.sort_values('date')\n",
    "df_20 = df_20.groupby('plot_number').apply(lambda x: x.iloc[2:]).reset_index(level=1, drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "id": "1ca1d33e-44d0-45b4-b333-0ce392deeef1",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_20grouped = df_20.groupby('plot_number')\n",
    "df_20long = df_20.copy()\n",
    "df_20long['obs_num'] = df_20grouped.cumcount() + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "id": "0473cfbe-7078-4f7c-b778-78c30bfed91f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "      <th>obs_num</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-06-28</td>\n",
       "      <td>0.017417</td>\n",
       "      <td>0.043442</td>\n",
       "      <td>0.020844</td>\n",
       "      <td>0.107552</td>\n",
       "      <td>0.433779</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-03</td>\n",
       "      <td>0.024159</td>\n",
       "      <td>0.055160</td>\n",
       "      <td>0.024431</td>\n",
       "      <td>0.131636</td>\n",
       "      <td>0.500632</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-08</td>\n",
       "      <td>0.022297</td>\n",
       "      <td>0.048837</td>\n",
       "      <td>0.020287</td>\n",
       "      <td>0.114728</td>\n",
       "      <td>0.444705</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-13</td>\n",
       "      <td>0.011046</td>\n",
       "      <td>0.031304</td>\n",
       "      <td>0.016580</td>\n",
       "      <td>0.070497</td>\n",
       "      <td>0.232059</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-18</td>\n",
       "      <td>0.022397</td>\n",
       "      <td>0.060377</td>\n",
       "      <td>0.048461</td>\n",
       "      <td>0.138072</td>\n",
       "      <td>0.290974</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-23</td>\n",
       "      <td>0.029142</td>\n",
       "      <td>0.077585</td>\n",
       "      <td>0.066680</td>\n",
       "      <td>0.177520</td>\n",
       "      <td>0.331992</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-07-28</td>\n",
       "      <td>0.027139</td>\n",
       "      <td>0.075271</td>\n",
       "      <td>0.070734</td>\n",
       "      <td>0.173504</td>\n",
       "      <td>0.311046</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-02</td>\n",
       "      <td>0.024164</td>\n",
       "      <td>0.068506</td>\n",
       "      <td>0.079378</td>\n",
       "      <td>0.161215</td>\n",
       "      <td>0.289961</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-07</td>\n",
       "      <td>0.031707</td>\n",
       "      <td>0.071970</td>\n",
       "      <td>0.110121</td>\n",
       "      <td>0.155774</td>\n",
       "      <td>0.253116</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2020-08-12</td>\n",
       "      <td>0.042600</td>\n",
       "      <td>0.077703</td>\n",
       "      <td>0.126265</td>\n",
       "      <td>0.151689</td>\n",
       "      <td>0.252579</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>2020-06-28</td>\n",
       "      <td>0.016551</td>\n",
       "      <td>0.042120</td>\n",
       "      <td>0.019487</td>\n",
       "      <td>0.107700</td>\n",
       "      <td>0.463768</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>2020-07-03</td>\n",
       "      <td>0.021390</td>\n",
       "      <td>0.049889</td>\n",
       "      <td>0.020438</td>\n",
       "      <td>0.128260</td>\n",
       "      <td>0.555080</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1105.0      2020-06-28     0.017417      0.043442    0.020844        0.107552   \n",
       "1105.0      2020-07-03     0.024159      0.055160    0.024431        0.131636   \n",
       "1105.0      2020-07-08     0.022297      0.048837    0.020287        0.114728   \n",
       "1105.0      2020-07-13     0.011046      0.031304    0.016580        0.070497   \n",
       "1105.0      2020-07-18     0.022397      0.060377    0.048461        0.138072   \n",
       "1105.0      2020-07-23     0.029142      0.077585    0.066680        0.177520   \n",
       "1105.0      2020-07-28     0.027139      0.075271    0.070734        0.173504   \n",
       "1105.0      2020-08-02     0.024164      0.068506    0.079378        0.161215   \n",
       "1105.0      2020-08-07     0.031707      0.071970    0.110121        0.155774   \n",
       "1105.0      2020-08-12     0.042600      0.077703    0.126265        0.151689   \n",
       "1106.0      2020-06-28     0.016551      0.042120    0.019487        0.107700   \n",
       "1106.0      2020-07-03     0.021390      0.049889    0.020438        0.128260   \n",
       "\n",
       "             nir_median  obs_num  \n",
       "plot_number                       \n",
       "1105.0         0.433779        1  \n",
       "1105.0         0.500632        2  \n",
       "1105.0         0.444705        3  \n",
       "1105.0         0.232059        4  \n",
       "1105.0         0.290974        5  \n",
       "1105.0         0.331992        6  \n",
       "1105.0         0.311046        7  \n",
       "1105.0         0.289961        8  \n",
       "1105.0         0.253116        9  \n",
       "1105.0         0.252579       10  \n",
       "1106.0         0.463768        1  \n",
       "1106.0         0.555080        2  "
      ]
     },
     "execution_count": 180,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_20long.head(12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "id": "4ddea80a-04b8-45a5-b253-7355a65826c8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"10\" halign=\"left\">blue_median</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"10\" halign=\"left\">rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>obs_num</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>...</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>0.017417</td>\n",
       "      <td>0.024159</td>\n",
       "      <td>0.022297</td>\n",
       "      <td>0.011046</td>\n",
       "      <td>0.022397</td>\n",
       "      <td>0.029142</td>\n",
       "      <td>0.027139</td>\n",
       "      <td>0.024164</td>\n",
       "      <td>0.031707</td>\n",
       "      <td>0.042600</td>\n",
       "      <td>...</td>\n",
       "      <td>0.107552</td>\n",
       "      <td>0.131636</td>\n",
       "      <td>0.114728</td>\n",
       "      <td>0.070497</td>\n",
       "      <td>0.138072</td>\n",
       "      <td>0.177520</td>\n",
       "      <td>0.173504</td>\n",
       "      <td>0.161215</td>\n",
       "      <td>0.155774</td>\n",
       "      <td>0.151689</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>0.016551</td>\n",
       "      <td>0.021390</td>\n",
       "      <td>0.019382</td>\n",
       "      <td>0.009555</td>\n",
       "      <td>0.023445</td>\n",
       "      <td>0.027993</td>\n",
       "      <td>0.024313</td>\n",
       "      <td>0.025539</td>\n",
       "      <td>0.035494</td>\n",
       "      <td>0.041878</td>\n",
       "      <td>...</td>\n",
       "      <td>0.107700</td>\n",
       "      <td>0.128260</td>\n",
       "      <td>0.107534</td>\n",
       "      <td>0.063179</td>\n",
       "      <td>0.133272</td>\n",
       "      <td>0.163817</td>\n",
       "      <td>0.161318</td>\n",
       "      <td>0.178479</td>\n",
       "      <td>0.188771</td>\n",
       "      <td>0.155710</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>0.017557</td>\n",
       "      <td>0.020047</td>\n",
       "      <td>0.016599</td>\n",
       "      <td>0.010219</td>\n",
       "      <td>0.024680</td>\n",
       "      <td>0.031687</td>\n",
       "      <td>0.028857</td>\n",
       "      <td>0.025150</td>\n",
       "      <td>0.030777</td>\n",
       "      <td>0.039861</td>\n",
       "      <td>...</td>\n",
       "      <td>0.101530</td>\n",
       "      <td>0.122705</td>\n",
       "      <td>0.104771</td>\n",
       "      <td>0.063426</td>\n",
       "      <td>0.156457</td>\n",
       "      <td>0.203933</td>\n",
       "      <td>0.195674</td>\n",
       "      <td>0.182954</td>\n",
       "      <td>0.174341</td>\n",
       "      <td>0.158242</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1109.0</th>\n",
       "      <td>0.018464</td>\n",
       "      <td>0.022135</td>\n",
       "      <td>0.019057</td>\n",
       "      <td>0.011995</td>\n",
       "      <td>0.028153</td>\n",
       "      <td>0.034959</td>\n",
       "      <td>0.029564</td>\n",
       "      <td>0.023003</td>\n",
       "      <td>0.031172</td>\n",
       "      <td>0.047361</td>\n",
       "      <td>...</td>\n",
       "      <td>0.111571</td>\n",
       "      <td>0.140801</td>\n",
       "      <td>0.124814</td>\n",
       "      <td>0.076021</td>\n",
       "      <td>0.154559</td>\n",
       "      <td>0.197366</td>\n",
       "      <td>0.186929</td>\n",
       "      <td>0.165611</td>\n",
       "      <td>0.156420</td>\n",
       "      <td>0.158969</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110.0</th>\n",
       "      <td>0.014360</td>\n",
       "      <td>0.016227</td>\n",
       "      <td>0.014689</td>\n",
       "      <td>0.009805</td>\n",
       "      <td>0.027209</td>\n",
       "      <td>0.032365</td>\n",
       "      <td>0.026652</td>\n",
       "      <td>0.023910</td>\n",
       "      <td>0.030340</td>\n",
       "      <td>0.037139</td>\n",
       "      <td>...</td>\n",
       "      <td>0.103380</td>\n",
       "      <td>0.125798</td>\n",
       "      <td>0.110801</td>\n",
       "      <td>0.071532</td>\n",
       "      <td>0.148859</td>\n",
       "      <td>0.207913</td>\n",
       "      <td>0.217904</td>\n",
       "      <td>0.198624</td>\n",
       "      <td>0.172742</td>\n",
       "      <td>0.148541</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1878.0</th>\n",
       "      <td>0.017929</td>\n",
       "      <td>0.022204</td>\n",
       "      <td>0.018173</td>\n",
       "      <td>0.011216</td>\n",
       "      <td>0.047829</td>\n",
       "      <td>0.049012</td>\n",
       "      <td>0.028031</td>\n",
       "      <td>0.022663</td>\n",
       "      <td>0.031074</td>\n",
       "      <td>0.035407</td>\n",
       "      <td>...</td>\n",
       "      <td>0.110245</td>\n",
       "      <td>0.136096</td>\n",
       "      <td>0.112525</td>\n",
       "      <td>0.070473</td>\n",
       "      <td>0.212552</td>\n",
       "      <td>0.243243</td>\n",
       "      <td>0.185349</td>\n",
       "      <td>0.155643</td>\n",
       "      <td>0.151129</td>\n",
       "      <td>0.141248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1879.0</th>\n",
       "      <td>0.016944</td>\n",
       "      <td>0.022300</td>\n",
       "      <td>0.018800</td>\n",
       "      <td>0.011347</td>\n",
       "      <td>0.049033</td>\n",
       "      <td>0.055386</td>\n",
       "      <td>0.041013</td>\n",
       "      <td>0.039936</td>\n",
       "      <td>0.047504</td>\n",
       "      <td>0.045170</td>\n",
       "      <td>...</td>\n",
       "      <td>0.100255</td>\n",
       "      <td>0.134212</td>\n",
       "      <td>0.118384</td>\n",
       "      <td>0.071283</td>\n",
       "      <td>0.206724</td>\n",
       "      <td>0.260137</td>\n",
       "      <td>0.234285</td>\n",
       "      <td>0.217017</td>\n",
       "      <td>0.199478</td>\n",
       "      <td>0.162057</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1880.0</th>\n",
       "      <td>0.017671</td>\n",
       "      <td>0.023194</td>\n",
       "      <td>0.018717</td>\n",
       "      <td>0.010951</td>\n",
       "      <td>0.046854</td>\n",
       "      <td>0.047009</td>\n",
       "      <td>0.025922</td>\n",
       "      <td>0.022087</td>\n",
       "      <td>0.030353</td>\n",
       "      <td>0.032876</td>\n",
       "      <td>...</td>\n",
       "      <td>0.105103</td>\n",
       "      <td>0.141638</td>\n",
       "      <td>0.119787</td>\n",
       "      <td>0.066310</td>\n",
       "      <td>0.191535</td>\n",
       "      <td>0.218317</td>\n",
       "      <td>0.168926</td>\n",
       "      <td>0.152347</td>\n",
       "      <td>0.151237</td>\n",
       "      <td>0.135903</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1881.0</th>\n",
       "      <td>0.016077</td>\n",
       "      <td>0.021689</td>\n",
       "      <td>0.018419</td>\n",
       "      <td>0.010978</td>\n",
       "      <td>0.045984</td>\n",
       "      <td>0.056090</td>\n",
       "      <td>0.050292</td>\n",
       "      <td>0.056861</td>\n",
       "      <td>0.063814</td>\n",
       "      <td>0.049098</td>\n",
       "      <td>...</td>\n",
       "      <td>0.098148</td>\n",
       "      <td>0.135390</td>\n",
       "      <td>0.116819</td>\n",
       "      <td>0.065637</td>\n",
       "      <td>0.200495</td>\n",
       "      <td>0.268069</td>\n",
       "      <td>0.264177</td>\n",
       "      <td>0.264033</td>\n",
       "      <td>0.244711</td>\n",
       "      <td>0.174751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1882.0</th>\n",
       "      <td>0.017935</td>\n",
       "      <td>0.022858</td>\n",
       "      <td>0.019427</td>\n",
       "      <td>0.011553</td>\n",
       "      <td>0.047428</td>\n",
       "      <td>0.050403</td>\n",
       "      <td>0.032484</td>\n",
       "      <td>0.029204</td>\n",
       "      <td>0.036369</td>\n",
       "      <td>0.039855</td>\n",
       "      <td>...</td>\n",
       "      <td>0.101564</td>\n",
       "      <td>0.133395</td>\n",
       "      <td>0.118961</td>\n",
       "      <td>0.068520</td>\n",
       "      <td>0.184919</td>\n",
       "      <td>0.240573</td>\n",
       "      <td>0.221625</td>\n",
       "      <td>0.192906</td>\n",
       "      <td>0.168244</td>\n",
       "      <td>0.148720</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>504 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            blue_median                                                    \\\n",
       "obs_num              1         2         3         4         5         6    \n",
       "plot_number                                                                 \n",
       "1105.0         0.017417  0.024159  0.022297  0.011046  0.022397  0.029142   \n",
       "1106.0         0.016551  0.021390  0.019382  0.009555  0.023445  0.027993   \n",
       "1108.0         0.017557  0.020047  0.016599  0.010219  0.024680  0.031687   \n",
       "1109.0         0.018464  0.022135  0.019057  0.011995  0.028153  0.034959   \n",
       "1110.0         0.014360  0.016227  0.014689  0.009805  0.027209  0.032365   \n",
       "...                 ...       ...       ...       ...       ...       ...   \n",
       "1878.0         0.017929  0.022204  0.018173  0.011216  0.047829  0.049012   \n",
       "1879.0         0.016944  0.022300  0.018800  0.011347  0.049033  0.055386   \n",
       "1880.0         0.017671  0.023194  0.018717  0.010951  0.046854  0.047009   \n",
       "1881.0         0.016077  0.021689  0.018419  0.010978  0.045984  0.056090   \n",
       "1882.0         0.017935  0.022858  0.019427  0.011553  0.047428  0.050403   \n",
       "\n",
       "                                                     ... rededge_median  \\\n",
       "obs_num            7         8         9         10  ...             1    \n",
       "plot_number                                          ...                  \n",
       "1105.0       0.027139  0.024164  0.031707  0.042600  ...       0.107552   \n",
       "1106.0       0.024313  0.025539  0.035494  0.041878  ...       0.107700   \n",
       "1108.0       0.028857  0.025150  0.030777  0.039861  ...       0.101530   \n",
       "1109.0       0.029564  0.023003  0.031172  0.047361  ...       0.111571   \n",
       "1110.0       0.026652  0.023910  0.030340  0.037139  ...       0.103380   \n",
       "...               ...       ...       ...       ...  ...            ...   \n",
       "1878.0       0.028031  0.022663  0.031074  0.035407  ...       0.110245   \n",
       "1879.0       0.041013  0.039936  0.047504  0.045170  ...       0.100255   \n",
       "1880.0       0.025922  0.022087  0.030353  0.032876  ...       0.105103   \n",
       "1881.0       0.050292  0.056861  0.063814  0.049098  ...       0.098148   \n",
       "1882.0       0.032484  0.029204  0.036369  0.039855  ...       0.101564   \n",
       "\n",
       "                                                                         \\\n",
       "obs_num            2         3         4         5         6         7    \n",
       "plot_number                                                               \n",
       "1105.0       0.131636  0.114728  0.070497  0.138072  0.177520  0.173504   \n",
       "1106.0       0.128260  0.107534  0.063179  0.133272  0.163817  0.161318   \n",
       "1108.0       0.122705  0.104771  0.063426  0.156457  0.203933  0.195674   \n",
       "1109.0       0.140801  0.124814  0.076021  0.154559  0.197366  0.186929   \n",
       "1110.0       0.125798  0.110801  0.071532  0.148859  0.207913  0.217904   \n",
       "...               ...       ...       ...       ...       ...       ...   \n",
       "1878.0       0.136096  0.112525  0.070473  0.212552  0.243243  0.185349   \n",
       "1879.0       0.134212  0.118384  0.071283  0.206724  0.260137  0.234285   \n",
       "1880.0       0.141638  0.119787  0.066310  0.191535  0.218317  0.168926   \n",
       "1881.0       0.135390  0.116819  0.065637  0.200495  0.268069  0.264177   \n",
       "1882.0       0.133395  0.118961  0.068520  0.184919  0.240573  0.221625   \n",
       "\n",
       "                                           \n",
       "obs_num            8         9         10  \n",
       "plot_number                                \n",
       "1105.0       0.161215  0.155774  0.151689  \n",
       "1106.0       0.178479  0.188771  0.155710  \n",
       "1108.0       0.182954  0.174341  0.158242  \n",
       "1109.0       0.165611  0.156420  0.158969  \n",
       "1110.0       0.198624  0.172742  0.148541  \n",
       "...               ...       ...       ...  \n",
       "1878.0       0.155643  0.151129  0.141248  \n",
       "1879.0       0.217017  0.199478  0.162057  \n",
       "1880.0       0.152347  0.151237  0.135903  \n",
       "1881.0       0.264033  0.244711  0.174751  \n",
       "1882.0       0.192906  0.168244  0.148720  \n",
       "\n",
       "[504 rows x 50 columns]"
      ]
     },
     "execution_count": 181,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_20pivot = df_20long.pivot_table(index='plot_number', columns='obs_num', values=['blue_median','green_median','red_median','rededge_median','nir_median'], aggfunc='mean')\n",
    "df_20pivot"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bf6ef816-e603-4f1f-9151-3620426ae775",
   "metadata": {},
   "source": [
    "## IQR for Outlier Detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "id": "e4f50c0f-92ef-4fe0-bd6c-a219909435c4",
   "metadata": {},
   "outputs": [],
   "source": [
    "q1 = np.percentile(df_20pivot, 20, axis=0)\n",
    "q3 = np.percentile(df_20pivot, 80, axis=0)\n",
    "iqr = q3 - q1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "id": "f15d2ced-3e32-425c-84b9-56c59e3f3696",
   "metadata": {},
   "outputs": [],
   "source": [
    "lower_bound = q1 - 1.5 * iqr\n",
    "upper_bound = q3 + 1.5 * iqr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "id": "97519bb8-72a0-4023-bd19-3a6b7129ac3a",
   "metadata": {},
   "outputs": [],
   "source": [
    "outliers = np.logical_or(df_20pivot < lower_bound, df_20pivot > upper_bound)\n",
    "df_20clean = df_20pivot[~outliers.any(axis=1)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "id": "6b01b066-0903-4695-a010-4d2d698ef37e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(422, 50)"
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_20clean.shape\n",
    "#Original Shape was (504, 50)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "05de0484-2165-4246-8c34-5096bd3479d6",
   "metadata": {},
   "source": [
    "#### Converting to Long Format again for Differencing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "id": "0ff72aee-2302-4d69-a23d-4270475bf138",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4220, 6)"
      ]
     },
     "execution_count": 196,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_20pivotc = df_20clean.copy()\n",
    "df_20pivotc.columns = [('X'.join(map(str, col))) for col in df_20pivot.columns.values]\n",
    "df_20wide = df_20pivotc.reset_index()\n",
    "df_20long = pd.melt(df_20wide, id_vars='plot_number', var_name='feature', value_name='value')\n",
    "df_20long[['color','time_step']] = df_20long['feature'].str.split('X', expand=True)\n",
    "df_20long = df_20long.drop('feature', axis=1)\n",
    "df_20long = df_20long.set_index(['plot_number', 'time_step', 'color'])\n",
    "\n",
    "df_20wide_new = df_20long.unstack(level=-1)\n",
    "df_20w = df_20wide_new.reset_index()\n",
    "\n",
    "df_20w.columns = [' '.join(map(str, col)).replace('_','_') for col in df_20w.columns.values]\n",
    "df_20w.columns.name = None\n",
    "\n",
    "df_20w['time_step '] = df_20w['time_step '].astype(int)\n",
    "df_20ws = df_20w.sort_values(by=['plot_number ','time_step '])\n",
    "\n",
    "df_20long = df_20ws.set_index('plot_number ')\n",
    "df_20long.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e3cb72bf-0d84-4df1-9e6b-7ecc85a3761f",
   "metadata": {},
   "source": [
    "##### Rest of the years are resampled as well before removing outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "id": "9d927975-9ab2-41bf-b7ad-9021dd2d7707",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(17280, 12)"
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X_2021all = df_X.loc[(df_X['season'] == 2021)]\n",
    "\n",
    "df_X_2021all.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "id": "aaee8dcd-ae15-46a1-a67a-b02b4522b3c3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12672, 12)"
      ]
     },
     "execution_count": 198,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X_2021 = df_X.loc[(df_X['season'] == 2021) & (df_X['camera'] == 'p4m')]\n",
    "\n",
    "df_X_2021.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "id": "16061c41-f152-4825-a73b-7af6910790af",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-05-28</td>\n",
       "      <td>0.028882</td>\n",
       "      <td>0.054156</td>\n",
       "      <td>0.047303</td>\n",
       "      <td>0.131320</td>\n",
       "      <td>0.162979</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-02</td>\n",
       "      <td>0.034943</td>\n",
       "      <td>0.068537</td>\n",
       "      <td>0.043628</td>\n",
       "      <td>0.218689</td>\n",
       "      <td>0.324878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-07</td>\n",
       "      <td>0.021654</td>\n",
       "      <td>0.040966</td>\n",
       "      <td>0.016993</td>\n",
       "      <td>0.195721</td>\n",
       "      <td>0.367871</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-12</td>\n",
       "      <td>0.020683</td>\n",
       "      <td>0.038784</td>\n",
       "      <td>0.017404</td>\n",
       "      <td>0.183363</td>\n",
       "      <td>0.355694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-17</td>\n",
       "      <td>0.020992</td>\n",
       "      <td>0.038800</td>\n",
       "      <td>0.019219</td>\n",
       "      <td>0.180083</td>\n",
       "      <td>0.353080</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-22</td>\n",
       "      <td>0.028040</td>\n",
       "      <td>0.043941</td>\n",
       "      <td>0.021129</td>\n",
       "      <td>0.204807</td>\n",
       "      <td>0.407503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-27</td>\n",
       "      <td>0.023153</td>\n",
       "      <td>0.044406</td>\n",
       "      <td>0.023317</td>\n",
       "      <td>0.199550</td>\n",
       "      <td>0.417437</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-02</td>\n",
       "      <td>0.017541</td>\n",
       "      <td>0.032946</td>\n",
       "      <td>0.017737</td>\n",
       "      <td>0.143223</td>\n",
       "      <td>0.293700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-07</td>\n",
       "      <td>0.000745</td>\n",
       "      <td>0.000929</td>\n",
       "      <td>0.000531</td>\n",
       "      <td>0.003213</td>\n",
       "      <td>0.006465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-12</td>\n",
       "      <td>0.024014</td>\n",
       "      <td>0.040452</td>\n",
       "      <td>0.024070</td>\n",
       "      <td>0.165522</td>\n",
       "      <td>0.303355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-17</td>\n",
       "      <td>0.019414</td>\n",
       "      <td>0.043275</td>\n",
       "      <td>0.032266</td>\n",
       "      <td>0.152824</td>\n",
       "      <td>0.255472</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-22</td>\n",
       "      <td>0.025865</td>\n",
       "      <td>0.056129</td>\n",
       "      <td>0.063722</td>\n",
       "      <td>0.163698</td>\n",
       "      <td>0.232042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-27</td>\n",
       "      <td>0.022912</td>\n",
       "      <td>0.043842</td>\n",
       "      <td>0.062572</td>\n",
       "      <td>0.108220</td>\n",
       "      <td>0.138038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-08-01</td>\n",
       "      <td>0.025790</td>\n",
       "      <td>0.054915</td>\n",
       "      <td>0.082208</td>\n",
       "      <td>0.126600</td>\n",
       "      <td>0.161584</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1107.0      2021-05-28     0.028882      0.054156    0.047303        0.131320   \n",
       "1107.0      2021-06-02     0.034943      0.068537    0.043628        0.218689   \n",
       "1107.0      2021-06-07     0.021654      0.040966    0.016993        0.195721   \n",
       "1107.0      2021-06-12     0.020683      0.038784    0.017404        0.183363   \n",
       "1107.0      2021-06-17     0.020992      0.038800    0.019219        0.180083   \n",
       "1107.0      2021-06-22     0.028040      0.043941    0.021129        0.204807   \n",
       "1107.0      2021-06-27     0.023153      0.044406    0.023317        0.199550   \n",
       "1107.0      2021-07-02     0.017541      0.032946    0.017737        0.143223   \n",
       "1107.0      2021-07-07     0.000745      0.000929    0.000531        0.003213   \n",
       "1107.0      2021-07-12     0.024014      0.040452    0.024070        0.165522   \n",
       "1107.0      2021-07-17     0.019414      0.043275    0.032266        0.152824   \n",
       "1107.0      2021-07-22     0.025865      0.056129    0.063722        0.163698   \n",
       "1107.0      2021-07-27     0.022912      0.043842    0.062572        0.108220   \n",
       "1107.0      2021-08-01     0.025790      0.054915    0.082208        0.126600   \n",
       "\n",
       "             nir_median  \n",
       "plot_number              \n",
       "1107.0         0.162979  \n",
       "1107.0         0.324878  \n",
       "1107.0         0.367871  \n",
       "1107.0         0.355694  \n",
       "1107.0         0.353080  \n",
       "1107.0         0.407503  \n",
       "1107.0         0.417437  \n",
       "1107.0         0.293700  \n",
       "1107.0         0.006465  \n",
       "1107.0         0.303355  \n",
       "1107.0         0.255472  \n",
       "1107.0         0.232042  \n",
       "1107.0         0.138038  \n",
       "1107.0         0.161584  "
      ]
     },
     "execution_count": 202,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21X = df_X_2021[['date','blue_median','green_median','red_median','rededge_median','nir_median']]\n",
    "df_21X.reset_index(inplace=True)\n",
    "\n",
    "freq = '5D'\n",
    "\n",
    "df_21res = pd.DataFrame()\n",
    "\n",
    "for plot_num in df_21X['plot_number'].unique():\n",
    "    \n",
    "    data = df_21X.loc[df_21X['plot_number'] == plot_num]\n",
    "    data.set_index('date', inplace = True)\n",
    "    data_res = data.resample(freq).mean().reset_index()\n",
    "    data_res['plot_number'] = plot_num\n",
    "    \n",
    "    df_21res = pd.concat([df_21res, data_res], ignore_index=False)\n",
    "\n",
    "df_21res.set_index('plot_number', inplace=True)\n",
    "df_21res.head(14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "id": "2049e1b1-f537-4b65-a252-c40a43e4dc1b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-05-28</td>\n",
       "      <td>0.028882</td>\n",
       "      <td>0.054156</td>\n",
       "      <td>0.047303</td>\n",
       "      <td>0.131320</td>\n",
       "      <td>0.162979</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-02</td>\n",
       "      <td>0.034943</td>\n",
       "      <td>0.068537</td>\n",
       "      <td>0.043628</td>\n",
       "      <td>0.218689</td>\n",
       "      <td>0.324878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-07</td>\n",
       "      <td>0.021654</td>\n",
       "      <td>0.040966</td>\n",
       "      <td>0.016993</td>\n",
       "      <td>0.195721</td>\n",
       "      <td>0.367871</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-12</td>\n",
       "      <td>0.020683</td>\n",
       "      <td>0.038784</td>\n",
       "      <td>0.017404</td>\n",
       "      <td>0.183363</td>\n",
       "      <td>0.355694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-17</td>\n",
       "      <td>0.020992</td>\n",
       "      <td>0.038800</td>\n",
       "      <td>0.019219</td>\n",
       "      <td>0.180083</td>\n",
       "      <td>0.353080</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-22</td>\n",
       "      <td>0.028040</td>\n",
       "      <td>0.043941</td>\n",
       "      <td>0.021129</td>\n",
       "      <td>0.204807</td>\n",
       "      <td>0.407503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-27</td>\n",
       "      <td>0.023153</td>\n",
       "      <td>0.044406</td>\n",
       "      <td>0.023317</td>\n",
       "      <td>0.199550</td>\n",
       "      <td>0.417437</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-02</td>\n",
       "      <td>0.017541</td>\n",
       "      <td>0.032946</td>\n",
       "      <td>0.017737</td>\n",
       "      <td>0.143223</td>\n",
       "      <td>0.293700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-07</td>\n",
       "      <td>0.000745</td>\n",
       "      <td>0.000929</td>\n",
       "      <td>0.000531</td>\n",
       "      <td>0.003213</td>\n",
       "      <td>0.006465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-12</td>\n",
       "      <td>0.024014</td>\n",
       "      <td>0.040452</td>\n",
       "      <td>0.024070</td>\n",
       "      <td>0.165522</td>\n",
       "      <td>0.303355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-17</td>\n",
       "      <td>0.019414</td>\n",
       "      <td>0.043275</td>\n",
       "      <td>0.032266</td>\n",
       "      <td>0.152824</td>\n",
       "      <td>0.255472</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-22</td>\n",
       "      <td>0.025865</td>\n",
       "      <td>0.056129</td>\n",
       "      <td>0.063722</td>\n",
       "      <td>0.163698</td>\n",
       "      <td>0.232042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-27</td>\n",
       "      <td>0.022912</td>\n",
       "      <td>0.043842</td>\n",
       "      <td>0.062572</td>\n",
       "      <td>0.108220</td>\n",
       "      <td>0.138038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-08-01</td>\n",
       "      <td>0.025790</td>\n",
       "      <td>0.054915</td>\n",
       "      <td>0.082208</td>\n",
       "      <td>0.126600</td>\n",
       "      <td>0.161584</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1107.0      2021-05-28     0.028882      0.054156    0.047303        0.131320   \n",
       "1107.0      2021-06-02     0.034943      0.068537    0.043628        0.218689   \n",
       "1107.0      2021-06-07     0.021654      0.040966    0.016993        0.195721   \n",
       "1107.0      2021-06-12     0.020683      0.038784    0.017404        0.183363   \n",
       "1107.0      2021-06-17     0.020992      0.038800    0.019219        0.180083   \n",
       "1107.0      2021-06-22     0.028040      0.043941    0.021129        0.204807   \n",
       "1107.0      2021-06-27     0.023153      0.044406    0.023317        0.199550   \n",
       "1107.0      2021-07-02     0.017541      0.032946    0.017737        0.143223   \n",
       "1107.0      2021-07-07     0.000745      0.000929    0.000531        0.003213   \n",
       "1107.0      2021-07-12     0.024014      0.040452    0.024070        0.165522   \n",
       "1107.0      2021-07-17     0.019414      0.043275    0.032266        0.152824   \n",
       "1107.0      2021-07-22     0.025865      0.056129    0.063722        0.163698   \n",
       "1107.0      2021-07-27     0.022912      0.043842    0.062572        0.108220   \n",
       "1107.0      2021-08-01     0.025790      0.054915    0.082208        0.126600   \n",
       "\n",
       "             nir_median  \n",
       "plot_number              \n",
       "1107.0         0.162979  \n",
       "1107.0         0.324878  \n",
       "1107.0         0.367871  \n",
       "1107.0         0.355694  \n",
       "1107.0         0.353080  \n",
       "1107.0         0.407503  \n",
       "1107.0         0.417437  \n",
       "1107.0         0.293700  \n",
       "1107.0         0.006465  \n",
       "1107.0         0.303355  \n",
       "1107.0         0.255472  \n",
       "1107.0         0.232042  \n",
       "1107.0         0.138038  \n",
       "1107.0         0.161584  "
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "for column in df_21res.columns:\n",
    "    if (column != 'date'):\n",
    "        mask = df_21res[column].isnull()  # Create a mask for missing values\n",
    "        x = np.arange(len(df_21res))\n",
    "        y = df_21res[column].values\n",
    "\n",
    "        # Create a cubic spline interpolation function\n",
    "        spline = CubicSpline(x[~mask], y[~mask])\n",
    "\n",
    "        # Fill missing values with interpolated values\n",
    "        df_21res.loc[mask, column] = spline(x[mask])\n",
    "\n",
    "\n",
    "#df_21res = df_21res.fillna(df_20res.median(axis=0))\n",
    "df_21res.head(14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "id": "db9e5420-398b-4cdb-962d-c0b7490bc1f3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-05-28</td>\n",
       "      <td>0.028882</td>\n",
       "      <td>0.054156</td>\n",
       "      <td>0.047303</td>\n",
       "      <td>0.131320</td>\n",
       "      <td>0.162979</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-02</td>\n",
       "      <td>0.034943</td>\n",
       "      <td>0.068537</td>\n",
       "      <td>0.043628</td>\n",
       "      <td>0.218689</td>\n",
       "      <td>0.324878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-07</td>\n",
       "      <td>0.021654</td>\n",
       "      <td>0.040966</td>\n",
       "      <td>0.016993</td>\n",
       "      <td>0.195721</td>\n",
       "      <td>0.367871</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-12</td>\n",
       "      <td>0.020683</td>\n",
       "      <td>0.038784</td>\n",
       "      <td>0.017404</td>\n",
       "      <td>0.183363</td>\n",
       "      <td>0.355694</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-17</td>\n",
       "      <td>0.020992</td>\n",
       "      <td>0.038800</td>\n",
       "      <td>0.019219</td>\n",
       "      <td>0.180083</td>\n",
       "      <td>0.353080</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1107.0      2021-05-28     0.028882      0.054156    0.047303        0.131320   \n",
       "1107.0      2021-06-02     0.034943      0.068537    0.043628        0.218689   \n",
       "1107.0      2021-06-07     0.021654      0.040966    0.016993        0.195721   \n",
       "1107.0      2021-06-12     0.020683      0.038784    0.017404        0.183363   \n",
       "1107.0      2021-06-17     0.020992      0.038800    0.019219        0.180083   \n",
       "\n",
       "             nir_median  \n",
       "plot_number              \n",
       "1107.0         0.162979  \n",
       "1107.0         0.324878  \n",
       "1107.0         0.367871  \n",
       "1107.0         0.355694  \n",
       "1107.0         0.353080  "
      ]
     },
     "execution_count": 204,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21res.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "id": "fbd3f5da-5f2a-493a-ad1e-91086aed69e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_21 = df_21res.copy()\n",
    "df_21 = df_21.sort_values('date')\n",
    "df_21 = df_21.groupby('plot_number').apply(lambda x: x.iloc[6:]).reset_index(level=1, drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "id": "63353306-091a-46c4-a5d0-34c28129b7e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_21grouped = df_21.groupby('plot_number')\n",
    "df_21long = df_21.copy()\n",
    "df_21long['obs_num'] = df_21grouped.cumcount() + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "id": "d1327fe3-7c32-40d6-add3-54b57cf9b0d4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "      <th>obs_num</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-06-27</td>\n",
       "      <td>0.023153</td>\n",
       "      <td>0.044406</td>\n",
       "      <td>0.023317</td>\n",
       "      <td>0.199550</td>\n",
       "      <td>0.417437</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-02</td>\n",
       "      <td>0.017541</td>\n",
       "      <td>0.032946</td>\n",
       "      <td>0.017737</td>\n",
       "      <td>0.143223</td>\n",
       "      <td>0.293700</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-07</td>\n",
       "      <td>0.000745</td>\n",
       "      <td>0.000929</td>\n",
       "      <td>0.000531</td>\n",
       "      <td>0.003213</td>\n",
       "      <td>0.006465</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-12</td>\n",
       "      <td>0.024014</td>\n",
       "      <td>0.040452</td>\n",
       "      <td>0.024070</td>\n",
       "      <td>0.165522</td>\n",
       "      <td>0.303355</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-17</td>\n",
       "      <td>0.019414</td>\n",
       "      <td>0.043275</td>\n",
       "      <td>0.032266</td>\n",
       "      <td>0.152824</td>\n",
       "      <td>0.255472</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-22</td>\n",
       "      <td>0.025865</td>\n",
       "      <td>0.056129</td>\n",
       "      <td>0.063722</td>\n",
       "      <td>0.163698</td>\n",
       "      <td>0.232042</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-07-27</td>\n",
       "      <td>0.022912</td>\n",
       "      <td>0.043842</td>\n",
       "      <td>0.062572</td>\n",
       "      <td>0.108220</td>\n",
       "      <td>0.138038</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-08-01</td>\n",
       "      <td>0.025790</td>\n",
       "      <td>0.054915</td>\n",
       "      <td>0.082208</td>\n",
       "      <td>0.126600</td>\n",
       "      <td>0.161584</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-08-06</td>\n",
       "      <td>0.024807</td>\n",
       "      <td>0.060752</td>\n",
       "      <td>0.095520</td>\n",
       "      <td>0.124921</td>\n",
       "      <td>0.138543</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2021-08-11</td>\n",
       "      <td>0.021380</td>\n",
       "      <td>0.053732</td>\n",
       "      <td>0.075984</td>\n",
       "      <td>0.111523</td>\n",
       "      <td>0.113024</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>2021-06-27</td>\n",
       "      <td>0.023423</td>\n",
       "      <td>0.043967</td>\n",
       "      <td>0.021831</td>\n",
       "      <td>0.209520</td>\n",
       "      <td>0.458970</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>2021-07-02</td>\n",
       "      <td>0.017983</td>\n",
       "      <td>0.032838</td>\n",
       "      <td>0.016807</td>\n",
       "      <td>0.150814</td>\n",
       "      <td>0.325007</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1107.0      2021-06-27     0.023153      0.044406    0.023317        0.199550   \n",
       "1107.0      2021-07-02     0.017541      0.032946    0.017737        0.143223   \n",
       "1107.0      2021-07-07     0.000745      0.000929    0.000531        0.003213   \n",
       "1107.0      2021-07-12     0.024014      0.040452    0.024070        0.165522   \n",
       "1107.0      2021-07-17     0.019414      0.043275    0.032266        0.152824   \n",
       "1107.0      2021-07-22     0.025865      0.056129    0.063722        0.163698   \n",
       "1107.0      2021-07-27     0.022912      0.043842    0.062572        0.108220   \n",
       "1107.0      2021-08-01     0.025790      0.054915    0.082208        0.126600   \n",
       "1107.0      2021-08-06     0.024807      0.060752    0.095520        0.124921   \n",
       "1107.0      2021-08-11     0.021380      0.053732    0.075984        0.111523   \n",
       "1108.0      2021-06-27     0.023423      0.043967    0.021831        0.209520   \n",
       "1108.0      2021-07-02     0.017983      0.032838    0.016807        0.150814   \n",
       "\n",
       "             nir_median  obs_num  \n",
       "plot_number                       \n",
       "1107.0         0.417437        1  \n",
       "1107.0         0.293700        2  \n",
       "1107.0         0.006465        3  \n",
       "1107.0         0.303355        4  \n",
       "1107.0         0.255472        5  \n",
       "1107.0         0.232042        6  \n",
       "1107.0         0.138038        7  \n",
       "1107.0         0.161584        8  \n",
       "1107.0         0.138543        9  \n",
       "1107.0         0.113024       10  \n",
       "1108.0         0.458970        1  \n",
       "1108.0         0.325007        2  "
      ]
     },
     "execution_count": 207,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21long.head(12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "id": "d9d6e1a2-b408-4da3-b5a6-f1e9bffefa84",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"10\" halign=\"left\">blue_median</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"10\" halign=\"left\">rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>obs_num</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>...</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>0.023153</td>\n",
       "      <td>0.017541</td>\n",
       "      <td>0.000745</td>\n",
       "      <td>0.024014</td>\n",
       "      <td>0.019414</td>\n",
       "      <td>0.025865</td>\n",
       "      <td>0.022912</td>\n",
       "      <td>0.025790</td>\n",
       "      <td>0.024807</td>\n",
       "      <td>0.021380</td>\n",
       "      <td>...</td>\n",
       "      <td>0.199550</td>\n",
       "      <td>0.143223</td>\n",
       "      <td>0.003213</td>\n",
       "      <td>0.165522</td>\n",
       "      <td>0.152824</td>\n",
       "      <td>0.163698</td>\n",
       "      <td>0.108220</td>\n",
       "      <td>0.126600</td>\n",
       "      <td>0.124921</td>\n",
       "      <td>0.111523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>0.023423</td>\n",
       "      <td>0.017983</td>\n",
       "      <td>0.000767</td>\n",
       "      <td>0.026138</td>\n",
       "      <td>0.020470</td>\n",
       "      <td>0.028221</td>\n",
       "      <td>0.025208</td>\n",
       "      <td>0.025974</td>\n",
       "      <td>0.029655</td>\n",
       "      <td>0.028834</td>\n",
       "      <td>...</td>\n",
       "      <td>0.209520</td>\n",
       "      <td>0.150814</td>\n",
       "      <td>0.003417</td>\n",
       "      <td>0.177548</td>\n",
       "      <td>0.163367</td>\n",
       "      <td>0.172333</td>\n",
       "      <td>0.114710</td>\n",
       "      <td>0.125495</td>\n",
       "      <td>0.140210</td>\n",
       "      <td>0.132314</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1109.0</th>\n",
       "      <td>0.022543</td>\n",
       "      <td>0.017939</td>\n",
       "      <td>0.000769</td>\n",
       "      <td>0.027834</td>\n",
       "      <td>0.020637</td>\n",
       "      <td>0.028171</td>\n",
       "      <td>0.024829</td>\n",
       "      <td>0.027838</td>\n",
       "      <td>0.033728</td>\n",
       "      <td>0.032021</td>\n",
       "      <td>...</td>\n",
       "      <td>0.188621</td>\n",
       "      <td>0.140692</td>\n",
       "      <td>0.003163</td>\n",
       "      <td>0.171654</td>\n",
       "      <td>0.154043</td>\n",
       "      <td>0.160658</td>\n",
       "      <td>0.115579</td>\n",
       "      <td>0.126553</td>\n",
       "      <td>0.143832</td>\n",
       "      <td>0.141349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110.0</th>\n",
       "      <td>0.022739</td>\n",
       "      <td>0.018070</td>\n",
       "      <td>0.000761</td>\n",
       "      <td>0.026355</td>\n",
       "      <td>0.019770</td>\n",
       "      <td>0.030068</td>\n",
       "      <td>0.025757</td>\n",
       "      <td>0.032079</td>\n",
       "      <td>0.035549</td>\n",
       "      <td>0.030546</td>\n",
       "      <td>...</td>\n",
       "      <td>0.210535</td>\n",
       "      <td>0.155380</td>\n",
       "      <td>0.003419</td>\n",
       "      <td>0.177391</td>\n",
       "      <td>0.160286</td>\n",
       "      <td>0.179731</td>\n",
       "      <td>0.131588</td>\n",
       "      <td>0.145924</td>\n",
       "      <td>0.157971</td>\n",
       "      <td>0.142160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1111.0</th>\n",
       "      <td>0.024480</td>\n",
       "      <td>0.020110</td>\n",
       "      <td>0.000832</td>\n",
       "      <td>0.028955</td>\n",
       "      <td>0.021820</td>\n",
       "      <td>0.032486</td>\n",
       "      <td>0.023825</td>\n",
       "      <td>0.027542</td>\n",
       "      <td>0.027122</td>\n",
       "      <td>0.022795</td>\n",
       "      <td>...</td>\n",
       "      <td>0.208086</td>\n",
       "      <td>0.159529</td>\n",
       "      <td>0.003573</td>\n",
       "      <td>0.187477</td>\n",
       "      <td>0.177087</td>\n",
       "      <td>0.186126</td>\n",
       "      <td>0.120310</td>\n",
       "      <td>0.128438</td>\n",
       "      <td>0.129145</td>\n",
       "      <td>0.115664</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1882.0</th>\n",
       "      <td>0.021258</td>\n",
       "      <td>0.018305</td>\n",
       "      <td>0.000816</td>\n",
       "      <td>0.035782</td>\n",
       "      <td>0.024161</td>\n",
       "      <td>0.028487</td>\n",
       "      <td>0.030639</td>\n",
       "      <td>0.031639</td>\n",
       "      <td>0.031310</td>\n",
       "      <td>0.027353</td>\n",
       "      <td>...</td>\n",
       "      <td>0.195243</td>\n",
       "      <td>0.148446</td>\n",
       "      <td>0.003624</td>\n",
       "      <td>0.199611</td>\n",
       "      <td>0.173233</td>\n",
       "      <td>0.193727</td>\n",
       "      <td>0.171511</td>\n",
       "      <td>0.167427</td>\n",
       "      <td>0.169790</td>\n",
       "      <td>0.152253</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1883.0</th>\n",
       "      <td>0.022936</td>\n",
       "      <td>0.018757</td>\n",
       "      <td>0.000809</td>\n",
       "      <td>0.032968</td>\n",
       "      <td>0.020882</td>\n",
       "      <td>0.023552</td>\n",
       "      <td>0.024804</td>\n",
       "      <td>0.025397</td>\n",
       "      <td>0.025356</td>\n",
       "      <td>0.022800</td>\n",
       "      <td>...</td>\n",
       "      <td>0.207469</td>\n",
       "      <td>0.153431</td>\n",
       "      <td>0.003647</td>\n",
       "      <td>0.200176</td>\n",
       "      <td>0.164425</td>\n",
       "      <td>0.174315</td>\n",
       "      <td>0.147263</td>\n",
       "      <td>0.142624</td>\n",
       "      <td>0.146284</td>\n",
       "      <td>0.133431</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1884.0</th>\n",
       "      <td>0.023377</td>\n",
       "      <td>0.019885</td>\n",
       "      <td>0.000804</td>\n",
       "      <td>0.031794</td>\n",
       "      <td>0.025492</td>\n",
       "      <td>0.027760</td>\n",
       "      <td>0.033166</td>\n",
       "      <td>0.040068</td>\n",
       "      <td>0.042988</td>\n",
       "      <td>0.035666</td>\n",
       "      <td>...</td>\n",
       "      <td>0.226042</td>\n",
       "      <td>0.171460</td>\n",
       "      <td>0.004009</td>\n",
       "      <td>0.208109</td>\n",
       "      <td>0.184474</td>\n",
       "      <td>0.182521</td>\n",
       "      <td>0.166312</td>\n",
       "      <td>0.178803</td>\n",
       "      <td>0.202382</td>\n",
       "      <td>0.180644</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1885.0</th>\n",
       "      <td>0.023292</td>\n",
       "      <td>0.018722</td>\n",
       "      <td>0.000787</td>\n",
       "      <td>0.029925</td>\n",
       "      <td>0.021827</td>\n",
       "      <td>0.024625</td>\n",
       "      <td>0.028554</td>\n",
       "      <td>0.028799</td>\n",
       "      <td>0.033135</td>\n",
       "      <td>0.032272</td>\n",
       "      <td>...</td>\n",
       "      <td>0.209726</td>\n",
       "      <td>0.152220</td>\n",
       "      <td>0.003498</td>\n",
       "      <td>0.190144</td>\n",
       "      <td>0.159610</td>\n",
       "      <td>0.167631</td>\n",
       "      <td>0.142316</td>\n",
       "      <td>0.141033</td>\n",
       "      <td>0.154617</td>\n",
       "      <td>0.151632</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1886.0</th>\n",
       "      <td>0.021373</td>\n",
       "      <td>0.017934</td>\n",
       "      <td>0.000787</td>\n",
       "      <td>0.030014</td>\n",
       "      <td>0.023136</td>\n",
       "      <td>0.025784</td>\n",
       "      <td>0.025767</td>\n",
       "      <td>0.026875</td>\n",
       "      <td>0.028423</td>\n",
       "      <td>0.025386</td>\n",
       "      <td>...</td>\n",
       "      <td>0.219100</td>\n",
       "      <td>0.163455</td>\n",
       "      <td>0.003948</td>\n",
       "      <td>0.224513</td>\n",
       "      <td>0.197772</td>\n",
       "      <td>0.200277</td>\n",
       "      <td>0.135780</td>\n",
       "      <td>0.142675</td>\n",
       "      <td>0.181834</td>\n",
       "      <td>0.136150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>576 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            blue_median                                                    \\\n",
       "obs_num              1         2         3         4         5         6    \n",
       "plot_number                                                                 \n",
       "1107.0         0.023153  0.017541  0.000745  0.024014  0.019414  0.025865   \n",
       "1108.0         0.023423  0.017983  0.000767  0.026138  0.020470  0.028221   \n",
       "1109.0         0.022543  0.017939  0.000769  0.027834  0.020637  0.028171   \n",
       "1110.0         0.022739  0.018070  0.000761  0.026355  0.019770  0.030068   \n",
       "1111.0         0.024480  0.020110  0.000832  0.028955  0.021820  0.032486   \n",
       "...                 ...       ...       ...       ...       ...       ...   \n",
       "1882.0         0.021258  0.018305  0.000816  0.035782  0.024161  0.028487   \n",
       "1883.0         0.022936  0.018757  0.000809  0.032968  0.020882  0.023552   \n",
       "1884.0         0.023377  0.019885  0.000804  0.031794  0.025492  0.027760   \n",
       "1885.0         0.023292  0.018722  0.000787  0.029925  0.021827  0.024625   \n",
       "1886.0         0.021373  0.017934  0.000787  0.030014  0.023136  0.025784   \n",
       "\n",
       "                                                     ... rededge_median  \\\n",
       "obs_num            7         8         9         10  ...             1    \n",
       "plot_number                                          ...                  \n",
       "1107.0       0.022912  0.025790  0.024807  0.021380  ...       0.199550   \n",
       "1108.0       0.025208  0.025974  0.029655  0.028834  ...       0.209520   \n",
       "1109.0       0.024829  0.027838  0.033728  0.032021  ...       0.188621   \n",
       "1110.0       0.025757  0.032079  0.035549  0.030546  ...       0.210535   \n",
       "1111.0       0.023825  0.027542  0.027122  0.022795  ...       0.208086   \n",
       "...               ...       ...       ...       ...  ...            ...   \n",
       "1882.0       0.030639  0.031639  0.031310  0.027353  ...       0.195243   \n",
       "1883.0       0.024804  0.025397  0.025356  0.022800  ...       0.207469   \n",
       "1884.0       0.033166  0.040068  0.042988  0.035666  ...       0.226042   \n",
       "1885.0       0.028554  0.028799  0.033135  0.032272  ...       0.209726   \n",
       "1886.0       0.025767  0.026875  0.028423  0.025386  ...       0.219100   \n",
       "\n",
       "                                                                         \\\n",
       "obs_num            2         3         4         5         6         7    \n",
       "plot_number                                                               \n",
       "1107.0       0.143223  0.003213  0.165522  0.152824  0.163698  0.108220   \n",
       "1108.0       0.150814  0.003417  0.177548  0.163367  0.172333  0.114710   \n",
       "1109.0       0.140692  0.003163  0.171654  0.154043  0.160658  0.115579   \n",
       "1110.0       0.155380  0.003419  0.177391  0.160286  0.179731  0.131588   \n",
       "1111.0       0.159529  0.003573  0.187477  0.177087  0.186126  0.120310   \n",
       "...               ...       ...       ...       ...       ...       ...   \n",
       "1882.0       0.148446  0.003624  0.199611  0.173233  0.193727  0.171511   \n",
       "1883.0       0.153431  0.003647  0.200176  0.164425  0.174315  0.147263   \n",
       "1884.0       0.171460  0.004009  0.208109  0.184474  0.182521  0.166312   \n",
       "1885.0       0.152220  0.003498  0.190144  0.159610  0.167631  0.142316   \n",
       "1886.0       0.163455  0.003948  0.224513  0.197772  0.200277  0.135780   \n",
       "\n",
       "                                           \n",
       "obs_num            8         9         10  \n",
       "plot_number                                \n",
       "1107.0       0.126600  0.124921  0.111523  \n",
       "1108.0       0.125495  0.140210  0.132314  \n",
       "1109.0       0.126553  0.143832  0.141349  \n",
       "1110.0       0.145924  0.157971  0.142160  \n",
       "1111.0       0.128438  0.129145  0.115664  \n",
       "...               ...       ...       ...  \n",
       "1882.0       0.167427  0.169790  0.152253  \n",
       "1883.0       0.142624  0.146284  0.133431  \n",
       "1884.0       0.178803  0.202382  0.180644  \n",
       "1885.0       0.141033  0.154617  0.151632  \n",
       "1886.0       0.142675  0.181834  0.136150  \n",
       "\n",
       "[576 rows x 50 columns]"
      ]
     },
     "execution_count": 209,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21pivot = df_21long.pivot_table(index='plot_number', columns='obs_num', values=['blue_median','green_median','red_median','rededge_median','nir_median'], aggfunc='mean')\n",
    "df_21pivot"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3170d784-e5d7-4b36-b5d6-30fbb7648fea",
   "metadata": {},
   "source": [
    "## IQR for Outlier Detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "de5176ae-043a-47bd-96bd-c009972e30de",
   "metadata": {},
   "outputs": [],
   "source": [
    "q1 = np.percentile(df_21pivot, 20, axis=0)\n",
    "q3 = np.percentile(df_21pivot, 80, axis=0)\n",
    "iqr = q3 - q1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "id": "a98e2b30-5125-4556-aa14-f458c4620c00",
   "metadata": {},
   "outputs": [],
   "source": [
    "lower_bound = q1 - 1.5 * iqr\n",
    "upper_bound = q3 + 1.5 * iqr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "id": "6a2bee28-1fcf-4a33-9d02-15023205b0dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "outliers = np.logical_or(df_21pivot < lower_bound, df_21pivot > upper_bound)\n",
    "df_21clean = df_21pivot[~outliers.any(axis=1)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "id": "8721429a-b16e-4e07-b62f-2c5d2451c3d6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(508, 50)"
      ]
     },
     "execution_count": 213,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21clean.shape\n",
    "#Original Shape was (576, 50)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "51ef044c-a0d1-4b61-a2a4-f281b5019f6e",
   "metadata": {},
   "source": [
    "#### Converting to Long Format again for Differencing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "id": "09442593-5bea-4956-a506-a1943194c713",
   "metadata": {},
   "outputs": [],
   "source": [
    "#df_21w.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "id": "18ab8f2e-b0cc-487c-81f6-d9afccab11d3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>time_step</th>\n",
       "      <th>value blue_median</th>\n",
       "      <th>value green_median</th>\n",
       "      <th>value nir_median</th>\n",
       "      <th>value red_median</th>\n",
       "      <th>value rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.023153</td>\n",
       "      <td>0.044406</td>\n",
       "      <td>0.417437</td>\n",
       "      <td>0.023317</td>\n",
       "      <td>0.199550</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2</td>\n",
       "      <td>0.017541</td>\n",
       "      <td>0.032946</td>\n",
       "      <td>0.293700</td>\n",
       "      <td>0.017737</td>\n",
       "      <td>0.143223</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>3</td>\n",
       "      <td>0.000745</td>\n",
       "      <td>0.000929</td>\n",
       "      <td>0.006465</td>\n",
       "      <td>0.000531</td>\n",
       "      <td>0.003213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>4</td>\n",
       "      <td>0.024014</td>\n",
       "      <td>0.040452</td>\n",
       "      <td>0.303355</td>\n",
       "      <td>0.024070</td>\n",
       "      <td>0.165522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>5</td>\n",
       "      <td>0.019414</td>\n",
       "      <td>0.043275</td>\n",
       "      <td>0.255472</td>\n",
       "      <td>0.032266</td>\n",
       "      <td>0.152824</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>6</td>\n",
       "      <td>0.025865</td>\n",
       "      <td>0.056129</td>\n",
       "      <td>0.232042</td>\n",
       "      <td>0.063722</td>\n",
       "      <td>0.163698</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>7</td>\n",
       "      <td>0.022912</td>\n",
       "      <td>0.043842</td>\n",
       "      <td>0.138038</td>\n",
       "      <td>0.062572</td>\n",
       "      <td>0.108220</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>8</td>\n",
       "      <td>0.025790</td>\n",
       "      <td>0.054915</td>\n",
       "      <td>0.161584</td>\n",
       "      <td>0.082208</td>\n",
       "      <td>0.126600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>9</td>\n",
       "      <td>0.024807</td>\n",
       "      <td>0.060752</td>\n",
       "      <td>0.138543</td>\n",
       "      <td>0.095520</td>\n",
       "      <td>0.124921</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>10</td>\n",
       "      <td>0.021380</td>\n",
       "      <td>0.053732</td>\n",
       "      <td>0.113024</td>\n",
       "      <td>0.075984</td>\n",
       "      <td>0.111523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.023423</td>\n",
       "      <td>0.043967</td>\n",
       "      <td>0.458970</td>\n",
       "      <td>0.021831</td>\n",
       "      <td>0.209520</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>2</td>\n",
       "      <td>0.017983</td>\n",
       "      <td>0.032838</td>\n",
       "      <td>0.325007</td>\n",
       "      <td>0.016807</td>\n",
       "      <td>0.150814</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              time_step   value blue_median  value green_median  \\\n",
       "plot_number                                                       \n",
       "1107.0                 1           0.023153            0.044406   \n",
       "1107.0                 2           0.017541            0.032946   \n",
       "1107.0                 3           0.000745            0.000929   \n",
       "1107.0                 4           0.024014            0.040452   \n",
       "1107.0                 5           0.019414            0.043275   \n",
       "1107.0                 6           0.025865            0.056129   \n",
       "1107.0                 7           0.022912            0.043842   \n",
       "1107.0                 8           0.025790            0.054915   \n",
       "1107.0                 9           0.024807            0.060752   \n",
       "1107.0                10           0.021380            0.053732   \n",
       "1108.0                 1           0.023423            0.043967   \n",
       "1108.0                 2           0.017983            0.032838   \n",
       "\n",
       "              value nir_median  value red_median  value rededge_median  \n",
       "plot_number                                                             \n",
       "1107.0                0.417437          0.023317              0.199550  \n",
       "1107.0                0.293700          0.017737              0.143223  \n",
       "1107.0                0.006465          0.000531              0.003213  \n",
       "1107.0                0.303355          0.024070              0.165522  \n",
       "1107.0                0.255472          0.032266              0.152824  \n",
       "1107.0                0.232042          0.063722              0.163698  \n",
       "1107.0                0.138038          0.062572              0.108220  \n",
       "1107.0                0.161584          0.082208              0.126600  \n",
       "1107.0                0.138543          0.095520              0.124921  \n",
       "1107.0                0.113024          0.075984              0.111523  \n",
       "1108.0                0.458970          0.021831              0.209520  \n",
       "1108.0                0.325007          0.016807              0.150814  "
      ]
     },
     "execution_count": 215,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21pivotc = df_21clean.copy()\n",
    "df_21pivotc.columns = [('X'.join(map(str, col))) for col in df_21pivot.columns.values]\n",
    "df_21wide = df_21pivotc.reset_index()\n",
    "df_21long = pd.melt(df_21wide, id_vars='plot_number', var_name='feature', value_name='value')\n",
    "df_21long[['color','time_step']] = df_21long['feature'].str.split('X', expand=True)\n",
    "df_21long = df_21long.drop('feature', axis=1)\n",
    "df_21long = df_21long.set_index(['plot_number', 'time_step', 'color'])\n",
    "\n",
    "df_21wide_new = df_21long.unstack(level=-1)\n",
    "df_21w = df_21wide_new.reset_index()\n",
    "\n",
    "df_21w.columns = [' '.join(map(str, col)).replace('_','_') for col in df_21w.columns.values]\n",
    "df_21w.columns.name = None\n",
    "\n",
    "\n",
    "df_21w['time_step '] = df_21w['time_step '].astype(int)\n",
    "df_21ws = df_21w.sort_values(by=['plot_number ','time_step '])\n",
    "\n",
    "df_21long = df_21ws.set_index('plot_number ')\n",
    "df_21long.head(12)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d560f6b-431c-4959-91ea-3c6b43c1040d",
   "metadata": {},
   "source": [
    "#### Repeating the same steps for 2022 Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "855e6880-e143-4e1f-8ac0-4520231b9b44",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "id": "ddb1cea2-4c2d-4fb2-b3eb-b9652da13d2a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(13824, 12)"
      ]
     },
     "execution_count": 216,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X_2022 = df_X.loc[(df_X['season'] == 2022)]\n",
    "\n",
    "df_X_2022.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "id": "ffcd3a7f-1592-41ee-a6bf-e8a1455102d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_22X = df_X_2022[['date','blue_median','green_median','red_median','rededge_median','nir_median']]\n",
    "df_22X.reset_index(inplace=True)\n",
    "\n",
    "freq = '5D'\n",
    "\n",
    "df_22res = pd.DataFrame()\n",
    "\n",
    "for plot_num in df_22X['plot_number'].unique():\n",
    "    \n",
    "    data = df_22X.loc[df_22X['plot_number'] == plot_num]\n",
    "    data.set_index('date', inplace = True)\n",
    "    data_res = data.resample(freq).mean().reset_index()\n",
    "    data_res['plot_number'] = plot_num\n",
    "    \n",
    "    df_22res = pd.concat([df_22res, data_res], ignore_index=False)\n",
    "\n",
    "df_22res.set_index('plot_number', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "id": "1af23b2c-8ada-4f0b-806d-09a5ddee25eb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-05-27</td>\n",
       "      <td>0.030947</td>\n",
       "      <td>0.053952</td>\n",
       "      <td>0.053469</td>\n",
       "      <td>0.103587</td>\n",
       "      <td>0.130319</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-06</td>\n",
       "      <td>0.030939</td>\n",
       "      <td>0.061026</td>\n",
       "      <td>0.037247</td>\n",
       "      <td>0.209947</td>\n",
       "      <td>0.325153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-11</td>\n",
       "      <td>0.020185</td>\n",
       "      <td>0.042952</td>\n",
       "      <td>0.021319</td>\n",
       "      <td>0.167530</td>\n",
       "      <td>0.293996</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-16</td>\n",
       "      <td>0.022985</td>\n",
       "      <td>0.043638</td>\n",
       "      <td>0.021716</td>\n",
       "      <td>0.177661</td>\n",
       "      <td>0.356251</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1107.0      2022-05-27     0.030947      0.053952    0.053469        0.103587   \n",
       "1107.0      2022-06-01          NaN           NaN         NaN             NaN   \n",
       "1107.0      2022-06-06     0.030939      0.061026    0.037247        0.209947   \n",
       "1107.0      2022-06-11     0.020185      0.042952    0.021319        0.167530   \n",
       "1107.0      2022-06-16     0.022985      0.043638    0.021716        0.177661   \n",
       "\n",
       "             nir_median  \n",
       "plot_number              \n",
       "1107.0         0.130319  \n",
       "1107.0              NaN  \n",
       "1107.0         0.325153  \n",
       "1107.0         0.293996  \n",
       "1107.0         0.356251  "
      ]
     },
     "execution_count": 218,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_22res.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "id": "3a35d322-784b-4b4a-8399-919c9ddfbbae",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-05-27</td>\n",
       "      <td>0.030947</td>\n",
       "      <td>0.053952</td>\n",
       "      <td>0.053469</td>\n",
       "      <td>0.103587</td>\n",
       "      <td>0.130319</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-01</td>\n",
       "      <td>0.042243</td>\n",
       "      <td>0.076363</td>\n",
       "      <td>0.055770</td>\n",
       "      <td>0.230307</td>\n",
       "      <td>0.341958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-06</td>\n",
       "      <td>0.030939</td>\n",
       "      <td>0.061026</td>\n",
       "      <td>0.037247</td>\n",
       "      <td>0.209947</td>\n",
       "      <td>0.325153</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-11</td>\n",
       "      <td>0.020185</td>\n",
       "      <td>0.042952</td>\n",
       "      <td>0.021319</td>\n",
       "      <td>0.167530</td>\n",
       "      <td>0.293996</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-16</td>\n",
       "      <td>0.022985</td>\n",
       "      <td>0.043638</td>\n",
       "      <td>0.021716</td>\n",
       "      <td>0.177661</td>\n",
       "      <td>0.356251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-21</td>\n",
       "      <td>0.021244</td>\n",
       "      <td>0.041135</td>\n",
       "      <td>0.020964</td>\n",
       "      <td>0.162244</td>\n",
       "      <td>0.319144</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-26</td>\n",
       "      <td>0.020501</td>\n",
       "      <td>0.039568</td>\n",
       "      <td>0.020830</td>\n",
       "      <td>0.162284</td>\n",
       "      <td>0.313354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-01</td>\n",
       "      <td>0.020667</td>\n",
       "      <td>0.042068</td>\n",
       "      <td>0.024877</td>\n",
       "      <td>0.170507</td>\n",
       "      <td>0.332424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-06</td>\n",
       "      <td>0.022994</td>\n",
       "      <td>0.048267</td>\n",
       "      <td>0.032257</td>\n",
       "      <td>0.185686</td>\n",
       "      <td>0.360964</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-11</td>\n",
       "      <td>0.027644</td>\n",
       "      <td>0.057879</td>\n",
       "      <td>0.042615</td>\n",
       "      <td>0.205316</td>\n",
       "      <td>0.383243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-16</td>\n",
       "      <td>0.028341</td>\n",
       "      <td>0.065849</td>\n",
       "      <td>0.054088</td>\n",
       "      <td>0.199414</td>\n",
       "      <td>0.350010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-21</td>\n",
       "      <td>0.028038</td>\n",
       "      <td>0.060129</td>\n",
       "      <td>0.053679</td>\n",
       "      <td>0.178914</td>\n",
       "      <td>0.305107</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-26</td>\n",
       "      <td>0.025718</td>\n",
       "      <td>0.063961</td>\n",
       "      <td>0.063569</td>\n",
       "      <td>0.174968</td>\n",
       "      <td>0.272090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-31</td>\n",
       "      <td>0.021840</td>\n",
       "      <td>0.051808</td>\n",
       "      <td>0.068687</td>\n",
       "      <td>0.141058</td>\n",
       "      <td>0.197469</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1107.0      2022-05-27     0.030947      0.053952    0.053469        0.103587   \n",
       "1107.0      2022-06-01     0.042243      0.076363    0.055770        0.230307   \n",
       "1107.0      2022-06-06     0.030939      0.061026    0.037247        0.209947   \n",
       "1107.0      2022-06-11     0.020185      0.042952    0.021319        0.167530   \n",
       "1107.0      2022-06-16     0.022985      0.043638    0.021716        0.177661   \n",
       "1107.0      2022-06-21     0.021244      0.041135    0.020964        0.162244   \n",
       "1107.0      2022-06-26     0.020501      0.039568    0.020830        0.162284   \n",
       "1107.0      2022-07-01     0.020667      0.042068    0.024877        0.170507   \n",
       "1107.0      2022-07-06     0.022994      0.048267    0.032257        0.185686   \n",
       "1107.0      2022-07-11     0.027644      0.057879    0.042615        0.205316   \n",
       "1107.0      2022-07-16     0.028341      0.065849    0.054088        0.199414   \n",
       "1107.0      2022-07-21     0.028038      0.060129    0.053679        0.178914   \n",
       "1107.0      2022-07-26     0.025718      0.063961    0.063569        0.174968   \n",
       "1107.0      2022-07-31     0.021840      0.051808    0.068687        0.141058   \n",
       "\n",
       "             nir_median  \n",
       "plot_number              \n",
       "1107.0         0.130319  \n",
       "1107.0         0.341958  \n",
       "1107.0         0.325153  \n",
       "1107.0         0.293996  \n",
       "1107.0         0.356251  \n",
       "1107.0         0.319144  \n",
       "1107.0         0.313354  \n",
       "1107.0         0.332424  \n",
       "1107.0         0.360964  \n",
       "1107.0         0.383243  \n",
       "1107.0         0.350010  \n",
       "1107.0         0.305107  \n",
       "1107.0         0.272090  \n",
       "1107.0         0.197469  "
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "for column in df_22res.columns:\n",
    "    if (column != 'date'):\n",
    "        mask = df_22res[column].isnull()  # Create a mask for missing values\n",
    "        x = np.arange(len(df_22res))\n",
    "        y = df_22res[column].values\n",
    "\n",
    "        # Create a cubic spline interpolation function\n",
    "        spline = CubicSpline(x[~mask], y[~mask])\n",
    "\n",
    "        # Fill missing values with interpolated values\n",
    "        df_22res.loc[mask, column] = spline(x[mask])\n",
    "\n",
    "#df_22res = df_22res.fillna(df_20res.median(axis=0))\n",
    "df_22res.head(14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "id": "42a26bf8-8443-4de4-922c-e4745523d788",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_22 = df_22res.copy()\n",
    "df_22 = df_22.sort_values('date')\n",
    "df_22 = df_22.groupby('plot_number').apply(lambda x: x.iloc[6:-3]).reset_index(level=1, drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "id": "7281d33d-4c42-440b-8381-e5a81b22d9c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_22grouped = df_22.groupby('plot_number')\n",
    "df_22long = df_22.copy()\n",
    "df_22long['obs_num'] = df_22grouped.cumcount() + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "id": "74e50f2f-263b-4d6d-b2bd-3034f5df335e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>date</th>\n",
       "      <th>blue_median</th>\n",
       "      <th>green_median</th>\n",
       "      <th>red_median</th>\n",
       "      <th>rededge_median</th>\n",
       "      <th>nir_median</th>\n",
       "      <th>obs_num</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-06-26</td>\n",
       "      <td>0.020501</td>\n",
       "      <td>0.039568</td>\n",
       "      <td>0.020830</td>\n",
       "      <td>0.162284</td>\n",
       "      <td>0.313354</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-01</td>\n",
       "      <td>0.020667</td>\n",
       "      <td>0.042068</td>\n",
       "      <td>0.024877</td>\n",
       "      <td>0.170507</td>\n",
       "      <td>0.332424</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-06</td>\n",
       "      <td>0.022994</td>\n",
       "      <td>0.048267</td>\n",
       "      <td>0.032257</td>\n",
       "      <td>0.185686</td>\n",
       "      <td>0.360964</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-11</td>\n",
       "      <td>0.027644</td>\n",
       "      <td>0.057879</td>\n",
       "      <td>0.042615</td>\n",
       "      <td>0.205316</td>\n",
       "      <td>0.383243</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-16</td>\n",
       "      <td>0.028341</td>\n",
       "      <td>0.065849</td>\n",
       "      <td>0.054088</td>\n",
       "      <td>0.199414</td>\n",
       "      <td>0.350010</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-21</td>\n",
       "      <td>0.028038</td>\n",
       "      <td>0.060129</td>\n",
       "      <td>0.053679</td>\n",
       "      <td>0.178914</td>\n",
       "      <td>0.305107</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-26</td>\n",
       "      <td>0.025718</td>\n",
       "      <td>0.063961</td>\n",
       "      <td>0.063569</td>\n",
       "      <td>0.174968</td>\n",
       "      <td>0.272090</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-07-31</td>\n",
       "      <td>0.021840</td>\n",
       "      <td>0.051808</td>\n",
       "      <td>0.068687</td>\n",
       "      <td>0.141058</td>\n",
       "      <td>0.197469</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-08-05</td>\n",
       "      <td>0.025443</td>\n",
       "      <td>0.052253</td>\n",
       "      <td>0.076677</td>\n",
       "      <td>0.129056</td>\n",
       "      <td>0.179857</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2022-08-10</td>\n",
       "      <td>0.030335</td>\n",
       "      <td>0.055948</td>\n",
       "      <td>0.089402</td>\n",
       "      <td>0.136189</td>\n",
       "      <td>0.190333</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>2022-06-26</td>\n",
       "      <td>0.023177</td>\n",
       "      <td>0.040688</td>\n",
       "      <td>0.018617</td>\n",
       "      <td>0.181775</td>\n",
       "      <td>0.374654</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>2022-07-01</td>\n",
       "      <td>0.025703</td>\n",
       "      <td>0.045514</td>\n",
       "      <td>0.022750</td>\n",
       "      <td>0.200025</td>\n",
       "      <td>0.434096</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  date  blue_median  green_median  red_median  rededge_median  \\\n",
       "plot_number                                                                     \n",
       "1107.0      2022-06-26     0.020501      0.039568    0.020830        0.162284   \n",
       "1107.0      2022-07-01     0.020667      0.042068    0.024877        0.170507   \n",
       "1107.0      2022-07-06     0.022994      0.048267    0.032257        0.185686   \n",
       "1107.0      2022-07-11     0.027644      0.057879    0.042615        0.205316   \n",
       "1107.0      2022-07-16     0.028341      0.065849    0.054088        0.199414   \n",
       "1107.0      2022-07-21     0.028038      0.060129    0.053679        0.178914   \n",
       "1107.0      2022-07-26     0.025718      0.063961    0.063569        0.174968   \n",
       "1107.0      2022-07-31     0.021840      0.051808    0.068687        0.141058   \n",
       "1107.0      2022-08-05     0.025443      0.052253    0.076677        0.129056   \n",
       "1107.0      2022-08-10     0.030335      0.055948    0.089402        0.136189   \n",
       "1108.0      2022-06-26     0.023177      0.040688    0.018617        0.181775   \n",
       "1108.0      2022-07-01     0.025703      0.045514    0.022750        0.200025   \n",
       "\n",
       "             nir_median  obs_num  \n",
       "plot_number                       \n",
       "1107.0         0.313354        1  \n",
       "1107.0         0.332424        2  \n",
       "1107.0         0.360964        3  \n",
       "1107.0         0.383243        4  \n",
       "1107.0         0.350010        5  \n",
       "1107.0         0.305107        6  \n",
       "1107.0         0.272090        7  \n",
       "1107.0         0.197469        8  \n",
       "1107.0         0.179857        9  \n",
       "1107.0         0.190333       10  \n",
       "1108.0         0.374654        1  \n",
       "1108.0         0.434096        2  "
      ]
     },
     "execution_count": 222,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_22long.head(12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "id": "55b72e57-72cf-4ae5-9958-e3569472204d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"10\" halign=\"left\">blue_median</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"10\" halign=\"left\">rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>obs_num</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>...</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>0.020501</td>\n",
       "      <td>0.020667</td>\n",
       "      <td>0.022994</td>\n",
       "      <td>0.027644</td>\n",
       "      <td>0.028341</td>\n",
       "      <td>0.028038</td>\n",
       "      <td>0.025718</td>\n",
       "      <td>0.021840</td>\n",
       "      <td>0.025443</td>\n",
       "      <td>0.030335</td>\n",
       "      <td>...</td>\n",
       "      <td>0.162284</td>\n",
       "      <td>0.170507</td>\n",
       "      <td>0.185686</td>\n",
       "      <td>0.205316</td>\n",
       "      <td>0.199414</td>\n",
       "      <td>0.178914</td>\n",
       "      <td>0.174968</td>\n",
       "      <td>0.141058</td>\n",
       "      <td>0.129056</td>\n",
       "      <td>0.136189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>0.023177</td>\n",
       "      <td>0.025703</td>\n",
       "      <td>0.029578</td>\n",
       "      <td>0.032562</td>\n",
       "      <td>0.032586</td>\n",
       "      <td>0.032713</td>\n",
       "      <td>0.032644</td>\n",
       "      <td>0.031812</td>\n",
       "      <td>0.036132</td>\n",
       "      <td>0.040255</td>\n",
       "      <td>...</td>\n",
       "      <td>0.181775</td>\n",
       "      <td>0.200025</td>\n",
       "      <td>0.218216</td>\n",
       "      <td>0.228984</td>\n",
       "      <td>0.221918</td>\n",
       "      <td>0.203109</td>\n",
       "      <td>0.204438</td>\n",
       "      <td>0.206949</td>\n",
       "      <td>0.180548</td>\n",
       "      <td>0.181229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1109.0</th>\n",
       "      <td>0.021538</td>\n",
       "      <td>0.023759</td>\n",
       "      <td>0.027219</td>\n",
       "      <td>0.031031</td>\n",
       "      <td>0.031863</td>\n",
       "      <td>0.032300</td>\n",
       "      <td>0.032597</td>\n",
       "      <td>0.032115</td>\n",
       "      <td>0.034457</td>\n",
       "      <td>0.036323</td>\n",
       "      <td>...</td>\n",
       "      <td>0.165612</td>\n",
       "      <td>0.183862</td>\n",
       "      <td>0.202430</td>\n",
       "      <td>0.220079</td>\n",
       "      <td>0.218391</td>\n",
       "      <td>0.200949</td>\n",
       "      <td>0.210415</td>\n",
       "      <td>0.214241</td>\n",
       "      <td>0.174929</td>\n",
       "      <td>0.166090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110.0</th>\n",
       "      <td>0.023095</td>\n",
       "      <td>0.025082</td>\n",
       "      <td>0.027869</td>\n",
       "      <td>0.030648</td>\n",
       "      <td>0.030781</td>\n",
       "      <td>0.030178</td>\n",
       "      <td>0.031460</td>\n",
       "      <td>0.028180</td>\n",
       "      <td>0.028993</td>\n",
       "      <td>0.034203</td>\n",
       "      <td>...</td>\n",
       "      <td>0.191758</td>\n",
       "      <td>0.211200</td>\n",
       "      <td>0.228912</td>\n",
       "      <td>0.236905</td>\n",
       "      <td>0.235234</td>\n",
       "      <td>0.211333</td>\n",
       "      <td>0.225427</td>\n",
       "      <td>0.209017</td>\n",
       "      <td>0.154209</td>\n",
       "      <td>0.166857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1111.0</th>\n",
       "      <td>0.025230</td>\n",
       "      <td>0.026623</td>\n",
       "      <td>0.029284</td>\n",
       "      <td>0.033947</td>\n",
       "      <td>0.035223</td>\n",
       "      <td>0.035290</td>\n",
       "      <td>0.033963</td>\n",
       "      <td>0.030269</td>\n",
       "      <td>0.033948</td>\n",
       "      <td>0.035704</td>\n",
       "      <td>...</td>\n",
       "      <td>0.160915</td>\n",
       "      <td>0.174958</td>\n",
       "      <td>0.193837</td>\n",
       "      <td>0.210789</td>\n",
       "      <td>0.209931</td>\n",
       "      <td>0.191456</td>\n",
       "      <td>0.187410</td>\n",
       "      <td>0.167032</td>\n",
       "      <td>0.140306</td>\n",
       "      <td>0.144401</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1882.0</th>\n",
       "      <td>0.022709</td>\n",
       "      <td>0.024630</td>\n",
       "      <td>0.027931</td>\n",
       "      <td>0.030883</td>\n",
       "      <td>0.029610</td>\n",
       "      <td>0.029800</td>\n",
       "      <td>0.031401</td>\n",
       "      <td>0.032576</td>\n",
       "      <td>0.032229</td>\n",
       "      <td>0.035005</td>\n",
       "      <td>...</td>\n",
       "      <td>0.170970</td>\n",
       "      <td>0.184458</td>\n",
       "      <td>0.206466</td>\n",
       "      <td>0.219708</td>\n",
       "      <td>0.207726</td>\n",
       "      <td>0.196833</td>\n",
       "      <td>0.212061</td>\n",
       "      <td>0.235089</td>\n",
       "      <td>0.179472</td>\n",
       "      <td>0.183492</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1883.0</th>\n",
       "      <td>0.024981</td>\n",
       "      <td>0.027423</td>\n",
       "      <td>0.030446</td>\n",
       "      <td>0.031998</td>\n",
       "      <td>0.031289</td>\n",
       "      <td>0.031947</td>\n",
       "      <td>0.031463</td>\n",
       "      <td>0.032662</td>\n",
       "      <td>0.033650</td>\n",
       "      <td>0.034654</td>\n",
       "      <td>...</td>\n",
       "      <td>0.189764</td>\n",
       "      <td>0.210847</td>\n",
       "      <td>0.232649</td>\n",
       "      <td>0.235090</td>\n",
       "      <td>0.228339</td>\n",
       "      <td>0.211776</td>\n",
       "      <td>0.220070</td>\n",
       "      <td>0.239097</td>\n",
       "      <td>0.177662</td>\n",
       "      <td>0.176170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1884.0</th>\n",
       "      <td>0.023872</td>\n",
       "      <td>0.024387</td>\n",
       "      <td>0.026269</td>\n",
       "      <td>0.028530</td>\n",
       "      <td>0.026696</td>\n",
       "      <td>0.026110</td>\n",
       "      <td>0.023012</td>\n",
       "      <td>0.022590</td>\n",
       "      <td>0.024182</td>\n",
       "      <td>0.027718</td>\n",
       "      <td>...</td>\n",
       "      <td>0.195421</td>\n",
       "      <td>0.207925</td>\n",
       "      <td>0.225632</td>\n",
       "      <td>0.232636</td>\n",
       "      <td>0.216516</td>\n",
       "      <td>0.196447</td>\n",
       "      <td>0.187558</td>\n",
       "      <td>0.205417</td>\n",
       "      <td>0.164557</td>\n",
       "      <td>0.171604</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1885.0</th>\n",
       "      <td>0.024939</td>\n",
       "      <td>0.025268</td>\n",
       "      <td>0.026868</td>\n",
       "      <td>0.028893</td>\n",
       "      <td>0.027703</td>\n",
       "      <td>0.027235</td>\n",
       "      <td>0.026699</td>\n",
       "      <td>0.029650</td>\n",
       "      <td>0.030213</td>\n",
       "      <td>0.032592</td>\n",
       "      <td>...</td>\n",
       "      <td>0.203441</td>\n",
       "      <td>0.217223</td>\n",
       "      <td>0.234097</td>\n",
       "      <td>0.238924</td>\n",
       "      <td>0.227111</td>\n",
       "      <td>0.206485</td>\n",
       "      <td>0.208198</td>\n",
       "      <td>0.231048</td>\n",
       "      <td>0.173488</td>\n",
       "      <td>0.175716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1886.0</th>\n",
       "      <td>0.022028</td>\n",
       "      <td>0.023738</td>\n",
       "      <td>0.024449</td>\n",
       "      <td>0.024516</td>\n",
       "      <td>0.022491</td>\n",
       "      <td>0.022304</td>\n",
       "      <td>0.021082</td>\n",
       "      <td>0.021478</td>\n",
       "      <td>0.024924</td>\n",
       "      <td>0.031200</td>\n",
       "      <td>...</td>\n",
       "      <td>0.197083</td>\n",
       "      <td>0.218570</td>\n",
       "      <td>0.229953</td>\n",
       "      <td>0.225146</td>\n",
       "      <td>0.204524</td>\n",
       "      <td>0.186619</td>\n",
       "      <td>0.187967</td>\n",
       "      <td>0.198277</td>\n",
       "      <td>0.175928</td>\n",
       "      <td>0.188811</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>576 rows × 50 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            blue_median                                                    \\\n",
       "obs_num              1         2         3         4         5         6    \n",
       "plot_number                                                                 \n",
       "1107.0         0.020501  0.020667  0.022994  0.027644  0.028341  0.028038   \n",
       "1108.0         0.023177  0.025703  0.029578  0.032562  0.032586  0.032713   \n",
       "1109.0         0.021538  0.023759  0.027219  0.031031  0.031863  0.032300   \n",
       "1110.0         0.023095  0.025082  0.027869  0.030648  0.030781  0.030178   \n",
       "1111.0         0.025230  0.026623  0.029284  0.033947  0.035223  0.035290   \n",
       "...                 ...       ...       ...       ...       ...       ...   \n",
       "1882.0         0.022709  0.024630  0.027931  0.030883  0.029610  0.029800   \n",
       "1883.0         0.024981  0.027423  0.030446  0.031998  0.031289  0.031947   \n",
       "1884.0         0.023872  0.024387  0.026269  0.028530  0.026696  0.026110   \n",
       "1885.0         0.024939  0.025268  0.026868  0.028893  0.027703  0.027235   \n",
       "1886.0         0.022028  0.023738  0.024449  0.024516  0.022491  0.022304   \n",
       "\n",
       "                                                     ... rededge_median  \\\n",
       "obs_num            7         8         9         10  ...             1    \n",
       "plot_number                                          ...                  \n",
       "1107.0       0.025718  0.021840  0.025443  0.030335  ...       0.162284   \n",
       "1108.0       0.032644  0.031812  0.036132  0.040255  ...       0.181775   \n",
       "1109.0       0.032597  0.032115  0.034457  0.036323  ...       0.165612   \n",
       "1110.0       0.031460  0.028180  0.028993  0.034203  ...       0.191758   \n",
       "1111.0       0.033963  0.030269  0.033948  0.035704  ...       0.160915   \n",
       "...               ...       ...       ...       ...  ...            ...   \n",
       "1882.0       0.031401  0.032576  0.032229  0.035005  ...       0.170970   \n",
       "1883.0       0.031463  0.032662  0.033650  0.034654  ...       0.189764   \n",
       "1884.0       0.023012  0.022590  0.024182  0.027718  ...       0.195421   \n",
       "1885.0       0.026699  0.029650  0.030213  0.032592  ...       0.203441   \n",
       "1886.0       0.021082  0.021478  0.024924  0.031200  ...       0.197083   \n",
       "\n",
       "                                                                         \\\n",
       "obs_num            2         3         4         5         6         7    \n",
       "plot_number                                                               \n",
       "1107.0       0.170507  0.185686  0.205316  0.199414  0.178914  0.174968   \n",
       "1108.0       0.200025  0.218216  0.228984  0.221918  0.203109  0.204438   \n",
       "1109.0       0.183862  0.202430  0.220079  0.218391  0.200949  0.210415   \n",
       "1110.0       0.211200  0.228912  0.236905  0.235234  0.211333  0.225427   \n",
       "1111.0       0.174958  0.193837  0.210789  0.209931  0.191456  0.187410   \n",
       "...               ...       ...       ...       ...       ...       ...   \n",
       "1882.0       0.184458  0.206466  0.219708  0.207726  0.196833  0.212061   \n",
       "1883.0       0.210847  0.232649  0.235090  0.228339  0.211776  0.220070   \n",
       "1884.0       0.207925  0.225632  0.232636  0.216516  0.196447  0.187558   \n",
       "1885.0       0.217223  0.234097  0.238924  0.227111  0.206485  0.208198   \n",
       "1886.0       0.218570  0.229953  0.225146  0.204524  0.186619  0.187967   \n",
       "\n",
       "                                           \n",
       "obs_num            8         9         10  \n",
       "plot_number                                \n",
       "1107.0       0.141058  0.129056  0.136189  \n",
       "1108.0       0.206949  0.180548  0.181229  \n",
       "1109.0       0.214241  0.174929  0.166090  \n",
       "1110.0       0.209017  0.154209  0.166857  \n",
       "1111.0       0.167032  0.140306  0.144401  \n",
       "...               ...       ...       ...  \n",
       "1882.0       0.235089  0.179472  0.183492  \n",
       "1883.0       0.239097  0.177662  0.176170  \n",
       "1884.0       0.205417  0.164557  0.171604  \n",
       "1885.0       0.231048  0.173488  0.175716  \n",
       "1886.0       0.198277  0.175928  0.188811  \n",
       "\n",
       "[576 rows x 50 columns]"
      ]
     },
     "execution_count": 223,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_22pivot = df_22long.pivot_table(index='plot_number', columns='obs_num', values=['blue_median','green_median','red_median','rededge_median','nir_median'], aggfunc='mean')\n",
    "df_22pivot"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3094f9c2-96a9-4a7c-9201-b5a00720d312",
   "metadata": {},
   "source": [
    "## IQR for Outlier Detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "id": "9de73cfa-3dd5-4855-a139-34ba38fd7b5d",
   "metadata": {},
   "outputs": [],
   "source": [
    "q1 = np.percentile(df_22pivot, 20, axis=0)\n",
    "q3 = np.percentile(df_22pivot, 80, axis=0)\n",
    "iqr = q3 - q1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "id": "0aa3697a-d8ed-4e5a-8cfa-4a4b94aaf6f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "lower_bound = q1 - 1.5 * iqr\n",
    "upper_bound = q3 + 1.5 * iqr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "id": "f38274f2-4823-4610-8318-be0c0b6cb861",
   "metadata": {},
   "outputs": [],
   "source": [
    "outliers = np.logical_or(df_22pivot < lower_bound, df_22pivot > upper_bound)\n",
    "df_22clean = df_22pivot[~outliers.any(axis=1)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "id": "462131c5-09f8-42d6-a9ce-83c9894cf39a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(513, 50)"
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_22clean.shape\n",
    "#Original Shape was (576, 50)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8b999b22-537b-4922-b241-f2ae5faf2910",
   "metadata": {},
   "source": [
    "#### Converting to Long Format again for Differencing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "id": "58593a66-2621-46d0-b5c7-9fc3370a93ee",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5130, 6)"
      ]
     },
     "execution_count": 228,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_22pivotc = df_22clean.copy()\n",
    "df_22pivotc.columns = [('X'.join(map(str, col))) for col in df_22pivot.columns.values]\n",
    "df_22wide = df_22pivotc.reset_index()\n",
    "df_22long = pd.melt(df_22wide, id_vars='plot_number', var_name='feature', value_name='value')\n",
    "df_22long[['color','time_step']] = df_22long['feature'].str.split('X', expand=True)\n",
    "df_22long = df_22long.drop('feature', axis=1)\n",
    "df_22long = df_22long.set_index(['plot_number', 'time_step', 'color'])\n",
    "\n",
    "df_22wide_new = df_22long.unstack(level=-1)\n",
    "df_22w = df_22wide_new.reset_index()\n",
    "\n",
    "df_22w.columns = [' '.join(map(str, col)).replace('_','_') for col in df_22w.columns.values]\n",
    "df_22w.columns.name = None\n",
    "\n",
    "\n",
    "df_22w['time_step '] = df_22w['time_step '].astype(int)\n",
    "df_22ws = df_22w.sort_values(by=['plot_number ','time_step '])\n",
    "\n",
    "df_22long = df_22ws.set_index('plot_number ')\n",
    "df_22long.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32463017-16eb-4350-ada4-7040a9f6b3f6",
   "metadata": {},
   "source": [
    "## Removing the Trend by differencing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "id": "2a6deee2-f3b9-4066-882f-916615bd28ed",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>time_step</th>\n",
       "      <th>value blue_median</th>\n",
       "      <th>value green_median</th>\n",
       "      <th>value nir_median</th>\n",
       "      <th>value red_median</th>\n",
       "      <th>value rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>3</td>\n",
       "      <td>-0.008605</td>\n",
       "      <td>-0.018040</td>\n",
       "      <td>-0.122780</td>\n",
       "      <td>-0.007731</td>\n",
       "      <td>-0.040992</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>4</td>\n",
       "      <td>-0.009388</td>\n",
       "      <td>-0.011211</td>\n",
       "      <td>-0.156719</td>\n",
       "      <td>0.000438</td>\n",
       "      <td>-0.027322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>5</td>\n",
       "      <td>0.022602</td>\n",
       "      <td>0.046606</td>\n",
       "      <td>0.271561</td>\n",
       "      <td>0.035587</td>\n",
       "      <td>0.111806</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>6</td>\n",
       "      <td>-0.004606</td>\n",
       "      <td>-0.011864</td>\n",
       "      <td>-0.017897</td>\n",
       "      <td>-0.013663</td>\n",
       "      <td>-0.028127</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>7</td>\n",
       "      <td>-0.008749</td>\n",
       "      <td>-0.019523</td>\n",
       "      <td>-0.061963</td>\n",
       "      <td>-0.014164</td>\n",
       "      <td>-0.043465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>8</td>\n",
       "      <td>-0.000972</td>\n",
       "      <td>-0.004451</td>\n",
       "      <td>-0.000139</td>\n",
       "      <td>0.004590</td>\n",
       "      <td>-0.008273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>9</td>\n",
       "      <td>0.010518</td>\n",
       "      <td>0.010230</td>\n",
       "      <td>-0.015761</td>\n",
       "      <td>0.022098</td>\n",
       "      <td>0.006849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>10</td>\n",
       "      <td>0.003351</td>\n",
       "      <td>0.002267</td>\n",
       "      <td>0.036309</td>\n",
       "      <td>-0.014598</td>\n",
       "      <td>0.001355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>1</td>\n",
       "      <td>-0.036944</td>\n",
       "      <td>-0.041314</td>\n",
       "      <td>0.211726</td>\n",
       "      <td>-0.122922</td>\n",
       "      <td>-0.039903</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              time_step   value blue_median  value green_median  \\\n",
       "plot_number                                                       \n",
       "1105.0                 1                NaN                 NaN   \n",
       "1105.0                 2                NaN                 NaN   \n",
       "1105.0                 3          -0.008605           -0.018040   \n",
       "1105.0                 4          -0.009388           -0.011211   \n",
       "1105.0                 5           0.022602            0.046606   \n",
       "1105.0                 6          -0.004606           -0.011864   \n",
       "1105.0                 7          -0.008749           -0.019523   \n",
       "1105.0                 8          -0.000972           -0.004451   \n",
       "1105.0                 9           0.010518            0.010230   \n",
       "1105.0                10           0.003351            0.002267   \n",
       "1106.0                 1          -0.036944           -0.041314   \n",
       "\n",
       "              value nir_median  value red_median  value rededge_median  \n",
       "plot_number                                                             \n",
       "1105.0                     NaN               NaN                   NaN  \n",
       "1105.0                     NaN               NaN                   NaN  \n",
       "1105.0               -0.122780         -0.007731             -0.040992  \n",
       "1105.0               -0.156719          0.000438             -0.027322  \n",
       "1105.0                0.271561          0.035587              0.111806  \n",
       "1105.0               -0.017897         -0.013663             -0.028127  \n",
       "1105.0               -0.061963         -0.014164             -0.043465  \n",
       "1105.0               -0.000139          0.004590             -0.008273  \n",
       "1105.0               -0.015761          0.022098              0.006849  \n",
       "1105.0                0.036309         -0.014598              0.001355  \n",
       "1106.0                0.211726         -0.122922             -0.039903  "
      ]
     },
     "execution_count": 232,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_20long_3f = df_20long.iloc[:,[0,1,2,3,4,5]]\n",
    "df_20diff1 = df_20long_3f.copy()\n",
    "df_20diff1.iloc[:, 1:] = df_20long_3f.iloc[:, 1:].diff()\n",
    "df_20diff = df_20diff1.copy()\n",
    "df_20diff.iloc[:, 1:] = df_20diff1.iloc[:, 1:].diff()\n",
    "df_20diff.head(11)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "848bcc3e-41c8-4055-8997-9bad4f005b7c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "id": "2bfc302d-178d-498d-bab7-12d1f9becf58",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>time_step</th>\n",
       "      <th>value blue_median</th>\n",
       "      <th>value green_median</th>\n",
       "      <th>value nir_median</th>\n",
       "      <th>value red_median</th>\n",
       "      <th>value rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>3</td>\n",
       "      <td>-0.011184</td>\n",
       "      <td>-0.020557</td>\n",
       "      <td>-0.163499</td>\n",
       "      <td>-0.011625</td>\n",
       "      <td>-0.083682</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>4</td>\n",
       "      <td>0.040064</td>\n",
       "      <td>0.071539</td>\n",
       "      <td>0.584125</td>\n",
       "      <td>0.040744</td>\n",
       "      <td>0.302318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>5</td>\n",
       "      <td>-0.027868</td>\n",
       "      <td>-0.036699</td>\n",
       "      <td>-0.344772</td>\n",
       "      <td>-0.015342</td>\n",
       "      <td>-0.175006</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>6</td>\n",
       "      <td>0.011051</td>\n",
       "      <td>0.010031</td>\n",
       "      <td>0.024453</td>\n",
       "      <td>0.023259</td>\n",
       "      <td>0.023571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>7</td>\n",
       "      <td>-0.009405</td>\n",
       "      <td>-0.025141</td>\n",
       "      <td>-0.070574</td>\n",
       "      <td>-0.032606</td>\n",
       "      <td>-0.066351</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>8</td>\n",
       "      <td>0.005832</td>\n",
       "      <td>0.023360</td>\n",
       "      <td>0.117550</td>\n",
       "      <td>0.020785</td>\n",
       "      <td>0.073858</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>9</td>\n",
       "      <td>-0.003862</td>\n",
       "      <td>-0.005237</td>\n",
       "      <td>-0.046587</td>\n",
       "      <td>-0.006323</td>\n",
       "      <td>-0.020059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>10</td>\n",
       "      <td>-0.002443</td>\n",
       "      <td>-0.012856</td>\n",
       "      <td>-0.002477</td>\n",
       "      <td>-0.032848</td>\n",
       "      <td>-0.011718</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.005470</td>\n",
       "      <td>-0.002745</td>\n",
       "      <td>0.371465</td>\n",
       "      <td>-0.034617</td>\n",
       "      <td>0.111394</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              time_step   value blue_median  value green_median  \\\n",
       "plot_number                                                       \n",
       "1107.0                 1                NaN                 NaN   \n",
       "1107.0                 2                NaN                 NaN   \n",
       "1107.0                 3          -0.011184           -0.020557   \n",
       "1107.0                 4           0.040064            0.071539   \n",
       "1107.0                 5          -0.027868           -0.036699   \n",
       "1107.0                 6           0.011051            0.010031   \n",
       "1107.0                 7          -0.009405           -0.025141   \n",
       "1107.0                 8           0.005832            0.023360   \n",
       "1107.0                 9          -0.003862           -0.005237   \n",
       "1107.0                10          -0.002443           -0.012856   \n",
       "1108.0                 1           0.005470           -0.002745   \n",
       "\n",
       "              value nir_median  value red_median  value rededge_median  \n",
       "plot_number                                                             \n",
       "1107.0                     NaN               NaN                   NaN  \n",
       "1107.0                     NaN               NaN                   NaN  \n",
       "1107.0               -0.163499         -0.011625             -0.083682  \n",
       "1107.0                0.584125          0.040744              0.302318  \n",
       "1107.0               -0.344772         -0.015342             -0.175006  \n",
       "1107.0                0.024453          0.023259              0.023571  \n",
       "1107.0               -0.070574         -0.032606             -0.066351  \n",
       "1107.0                0.117550          0.020785              0.073858  \n",
       "1107.0               -0.046587         -0.006323             -0.020059  \n",
       "1107.0               -0.002477         -0.032848             -0.011718  \n",
       "1108.0                0.371465         -0.034617              0.111394  "
      ]
     },
     "execution_count": 233,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21long_3f = df_21long.iloc[:,[0,1,2,3,4,5]]\n",
    "df_21diff1 = df_21long_3f.copy()\n",
    "df_21diff1.iloc[:, 1:] = df_21long_3f.iloc[:, 1:].diff()\n",
    "\n",
    "df_21diff = df_21diff1.copy()\n",
    "df_21diff.iloc[:, 1:] = df_21diff1.iloc[:, 1:].diff()\n",
    "df_21diff.head(11)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "id": "982e33ee-3e79-46c7-803a-c65000fedf62",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>time_step</th>\n",
       "      <th>value blue_median</th>\n",
       "      <th>value green_median</th>\n",
       "      <th>value nir_median</th>\n",
       "      <th>value red_median</th>\n",
       "      <th>value rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>3</td>\n",
       "      <td>0.001350</td>\n",
       "      <td>0.001199</td>\n",
       "      <td>-0.008880</td>\n",
       "      <td>0.001443</td>\n",
       "      <td>-0.000059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>4</td>\n",
       "      <td>-0.000892</td>\n",
       "      <td>-0.002318</td>\n",
       "      <td>-0.045976</td>\n",
       "      <td>-0.001265</td>\n",
       "      <td>-0.007422</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>5</td>\n",
       "      <td>-0.002960</td>\n",
       "      <td>0.004056</td>\n",
       "      <td>-0.045406</td>\n",
       "      <td>0.004786</td>\n",
       "      <td>-0.017835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>6</td>\n",
       "      <td>0.000103</td>\n",
       "      <td>-0.010730</td>\n",
       "      <td>-0.010826</td>\n",
       "      <td>-0.007648</td>\n",
       "      <td>-0.011743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>7</td>\n",
       "      <td>-0.000196</td>\n",
       "      <td>0.009866</td>\n",
       "      <td>0.019947</td>\n",
       "      <td>0.007071</td>\n",
       "      <td>0.020138</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>8</td>\n",
       "      <td>-0.000763</td>\n",
       "      <td>0.000685</td>\n",
       "      <td>-0.047357</td>\n",
       "      <td>0.020039</td>\n",
       "      <td>0.001182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>9</td>\n",
       "      <td>0.005153</td>\n",
       "      <td>-0.007870</td>\n",
       "      <td>0.029637</td>\n",
       "      <td>-0.004761</td>\n",
       "      <td>-0.028912</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>10</td>\n",
       "      <td>-0.000198</td>\n",
       "      <td>0.003585</td>\n",
       "      <td>0.051732</td>\n",
       "      <td>-0.003791</td>\n",
       "      <td>0.027081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1109.0</th>\n",
       "      <td>1</td>\n",
       "      <td>-0.022839</td>\n",
       "      <td>-0.043217</td>\n",
       "      <td>0.097488</td>\n",
       "      <td>-0.126219</td>\n",
       "      <td>-0.016297</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              time_step   value blue_median  value green_median  \\\n",
       "plot_number                                                       \n",
       "1108.0                 1                NaN                 NaN   \n",
       "1108.0                 2                NaN                 NaN   \n",
       "1108.0                 3           0.001350            0.001199   \n",
       "1108.0                 4          -0.000892           -0.002318   \n",
       "1108.0                 5          -0.002960            0.004056   \n",
       "1108.0                 6           0.000103           -0.010730   \n",
       "1108.0                 7          -0.000196            0.009866   \n",
       "1108.0                 8          -0.000763            0.000685   \n",
       "1108.0                 9           0.005153           -0.007870   \n",
       "1108.0                10          -0.000198            0.003585   \n",
       "1109.0                 1          -0.022839           -0.043217   \n",
       "\n",
       "              value nir_median  value red_median  value rededge_median  \n",
       "plot_number                                                             \n",
       "1108.0                     NaN               NaN                   NaN  \n",
       "1108.0                     NaN               NaN                   NaN  \n",
       "1108.0               -0.008880          0.001443             -0.000059  \n",
       "1108.0               -0.045976         -0.001265             -0.007422  \n",
       "1108.0               -0.045406          0.004786             -0.017835  \n",
       "1108.0               -0.010826         -0.007648             -0.011743  \n",
       "1108.0                0.019947          0.007071              0.020138  \n",
       "1108.0               -0.047357          0.020039              0.001182  \n",
       "1108.0                0.029637         -0.004761             -0.028912  \n",
       "1108.0                0.051732         -0.003791              0.027081  \n",
       "1109.0                0.097488         -0.126219             -0.016297  "
      ]
     },
     "execution_count": 234,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_22long_3f = df_22long.iloc[:,[0,1,2,3,4,5]]\n",
    "df_22diff1 = df_22long_3f.copy()\n",
    "df_22diff1.iloc[:, 1:] = df_22long_3f.iloc[:, 1:].diff()\n",
    "df_22diff = df_22diff1.copy()\n",
    "df_22diff.iloc[:, 1:] = df_22diff1.iloc[:, 1:].diff()\n",
    "df_22diff.head(11)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "id": "35885fdf-526b-4b69-be5c-966644625eb3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"6\" halign=\"left\">value blue_median</th>\n",
       "      <th colspan=\"4\" halign=\"left\">value green_median</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"4\" halign=\"left\">value red_median</th>\n",
       "      <th colspan=\"6\" halign=\"left\">value rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>time_step</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>...</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>-0.008605</td>\n",
       "      <td>-0.009388</td>\n",
       "      <td>0.022602</td>\n",
       "      <td>-0.004606</td>\n",
       "      <td>-0.008749</td>\n",
       "      <td>-0.000972</td>\n",
       "      <td>-0.018040</td>\n",
       "      <td>-0.011211</td>\n",
       "      <td>0.046606</td>\n",
       "      <td>-0.011864</td>\n",
       "      <td>...</td>\n",
       "      <td>0.035587</td>\n",
       "      <td>-0.013663</td>\n",
       "      <td>-0.014164</td>\n",
       "      <td>0.004590</td>\n",
       "      <td>-0.040992</td>\n",
       "      <td>-0.027322</td>\n",
       "      <td>0.111806</td>\n",
       "      <td>-0.028127</td>\n",
       "      <td>-0.043465</td>\n",
       "      <td>-0.008273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>-0.006848</td>\n",
       "      <td>-0.007819</td>\n",
       "      <td>0.023717</td>\n",
       "      <td>-0.009342</td>\n",
       "      <td>-0.008228</td>\n",
       "      <td>0.004907</td>\n",
       "      <td>-0.015791</td>\n",
       "      <td>-0.007327</td>\n",
       "      <td>0.049493</td>\n",
       "      <td>-0.021909</td>\n",
       "      <td>...</td>\n",
       "      <td>0.033554</td>\n",
       "      <td>-0.017280</td>\n",
       "      <td>-0.008371</td>\n",
       "      <td>0.017430</td>\n",
       "      <td>-0.041287</td>\n",
       "      <td>-0.023628</td>\n",
       "      <td>0.114447</td>\n",
       "      <td>-0.039549</td>\n",
       "      <td>-0.033043</td>\n",
       "      <td>0.019659</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>-0.005938</td>\n",
       "      <td>-0.002933</td>\n",
       "      <td>0.020841</td>\n",
       "      <td>-0.007454</td>\n",
       "      <td>-0.009837</td>\n",
       "      <td>-0.000877</td>\n",
       "      <td>-0.015265</td>\n",
       "      <td>-0.004520</td>\n",
       "      <td>0.058495</td>\n",
       "      <td>-0.026264</td>\n",
       "      <td>...</td>\n",
       "      <td>0.046000</td>\n",
       "      <td>-0.020196</td>\n",
       "      <td>-0.017759</td>\n",
       "      <td>0.008623</td>\n",
       "      <td>-0.039108</td>\n",
       "      <td>-0.023412</td>\n",
       "      <td>0.134376</td>\n",
       "      <td>-0.045555</td>\n",
       "      <td>-0.055735</td>\n",
       "      <td>-0.004461</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1109.0</th>\n",
       "      <td>-0.006748</td>\n",
       "      <td>-0.003985</td>\n",
       "      <td>0.023220</td>\n",
       "      <td>-0.009352</td>\n",
       "      <td>-0.012201</td>\n",
       "      <td>-0.001166</td>\n",
       "      <td>-0.017631</td>\n",
       "      <td>-0.007990</td>\n",
       "      <td>0.056392</td>\n",
       "      <td>-0.023796</td>\n",
       "      <td>...</td>\n",
       "      <td>0.039091</td>\n",
       "      <td>-0.011797</td>\n",
       "      <td>-0.018655</td>\n",
       "      <td>-0.002589</td>\n",
       "      <td>-0.045216</td>\n",
       "      <td>-0.032807</td>\n",
       "      <td>0.127332</td>\n",
       "      <td>-0.035732</td>\n",
       "      <td>-0.053243</td>\n",
       "      <td>-0.010882</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110.0</th>\n",
       "      <td>-0.003405</td>\n",
       "      <td>-0.003346</td>\n",
       "      <td>0.022287</td>\n",
       "      <td>-0.012247</td>\n",
       "      <td>-0.010869</td>\n",
       "      <td>0.002972</td>\n",
       "      <td>-0.013574</td>\n",
       "      <td>-0.005082</td>\n",
       "      <td>0.051582</td>\n",
       "      <td>-0.016651</td>\n",
       "      <td>...</td>\n",
       "      <td>0.034704</td>\n",
       "      <td>-0.002846</td>\n",
       "      <td>-0.010452</td>\n",
       "      <td>-0.003176</td>\n",
       "      <td>-0.037416</td>\n",
       "      <td>-0.024271</td>\n",
       "      <td>0.116596</td>\n",
       "      <td>-0.018274</td>\n",
       "      <td>-0.049061</td>\n",
       "      <td>-0.029272</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             value blue_median                                          \\\n",
       "time_step                    3         4         5         6         7   \n",
       "plot_number                                                              \n",
       "1105.0               -0.008605 -0.009388  0.022602 -0.004606 -0.008749   \n",
       "1106.0               -0.006848 -0.007819  0.023717 -0.009342 -0.008228   \n",
       "1108.0               -0.005938 -0.002933  0.020841 -0.007454 -0.009837   \n",
       "1109.0               -0.006748 -0.003985  0.023220 -0.009352 -0.012201   \n",
       "1110.0               -0.003405 -0.003346  0.022287 -0.012247 -0.010869   \n",
       "\n",
       "                       value green_median                                ...  \\\n",
       "time_step            8                  3         4         5         6  ...   \n",
       "plot_number                                                              ...   \n",
       "1105.0       -0.000972          -0.018040 -0.011211  0.046606 -0.011864  ...   \n",
       "1106.0        0.004907          -0.015791 -0.007327  0.049493 -0.021909  ...   \n",
       "1108.0       -0.000877          -0.015265 -0.004520  0.058495 -0.026264  ...   \n",
       "1109.0       -0.001166          -0.017631 -0.007990  0.056392 -0.023796  ...   \n",
       "1110.0        0.002972          -0.013574 -0.005082  0.051582 -0.016651  ...   \n",
       "\n",
       "             value red_median                                \\\n",
       "time_step                   5         6         7         8   \n",
       "plot_number                                                   \n",
       "1105.0               0.035587 -0.013663 -0.014164  0.004590   \n",
       "1106.0               0.033554 -0.017280 -0.008371  0.017430   \n",
       "1108.0               0.046000 -0.020196 -0.017759  0.008623   \n",
       "1109.0               0.039091 -0.011797 -0.018655 -0.002589   \n",
       "1110.0               0.034704 -0.002846 -0.010452 -0.003176   \n",
       "\n",
       "             value rededge_median                                          \\\n",
       "time_step                       3         4         5         6         7   \n",
       "plot_number                                                                 \n",
       "1105.0                  -0.040992 -0.027322  0.111806 -0.028127 -0.043465   \n",
       "1106.0                  -0.041287 -0.023628  0.114447 -0.039549 -0.033043   \n",
       "1108.0                  -0.039108 -0.023412  0.134376 -0.045555 -0.055735   \n",
       "1109.0                  -0.045216 -0.032807  0.127332 -0.035732 -0.053243   \n",
       "1110.0                  -0.037416 -0.024271  0.116596 -0.018274 -0.049061   \n",
       "\n",
       "                        \n",
       "time_step            8  \n",
       "plot_number             \n",
       "1105.0       -0.008273  \n",
       "1106.0        0.019659  \n",
       "1108.0       -0.004461  \n",
       "1109.0       -0.010882  \n",
       "1110.0       -0.029272  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 235,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Converting to wide format\n",
    "#The sequence of features is changed using this method but it is used to avoid an error\n",
    "\n",
    "df_20pf = df_20diff.pivot_table(index='plot_number ', columns='time_step ', values=['value blue_median','value green_median','value red_median','value rededge_median','value nir_median'], aggfunc='mean')\n",
    "\n",
    "columns_to_remove = [0,1,8,9,10,11,18,19,20,21,28,29,30,31,38,39,40,41,48,49]\n",
    "\n",
    "# Remove first and last observation from each plot\n",
    "df_20p = df_20pf.iloc[:, [i for i in range(df_20pf.shape[1]) if i not in columns_to_remove]]\n",
    "df_20p.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "id": "790fc0cc-f7f9-472e-a64b-64c132db4fd3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"6\" halign=\"left\">value blue_median</th>\n",
       "      <th colspan=\"4\" halign=\"left\">value green_median</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"4\" halign=\"left\">value red_median</th>\n",
       "      <th colspan=\"6\" halign=\"left\">value rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>time_step</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>...</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1107.0</th>\n",
       "      <td>-0.011184</td>\n",
       "      <td>0.040064</td>\n",
       "      <td>-0.027868</td>\n",
       "      <td>0.011051</td>\n",
       "      <td>-0.009405</td>\n",
       "      <td>0.005832</td>\n",
       "      <td>-0.020557</td>\n",
       "      <td>0.071539</td>\n",
       "      <td>-0.036699</td>\n",
       "      <td>0.010031</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.015342</td>\n",
       "      <td>0.023259</td>\n",
       "      <td>-0.032606</td>\n",
       "      <td>0.020785</td>\n",
       "      <td>-0.083682</td>\n",
       "      <td>0.302318</td>\n",
       "      <td>-0.175006</td>\n",
       "      <td>0.023571</td>\n",
       "      <td>-0.066351</td>\n",
       "      <td>0.073858</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>-0.011775</td>\n",
       "      <td>0.042587</td>\n",
       "      <td>-0.031039</td>\n",
       "      <td>0.013420</td>\n",
       "      <td>-0.010765</td>\n",
       "      <td>0.003780</td>\n",
       "      <td>-0.020754</td>\n",
       "      <td>0.073087</td>\n",
       "      <td>-0.037642</td>\n",
       "      <td>0.008480</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.013807</td>\n",
       "      <td>0.025891</td>\n",
       "      <td>-0.033949</td>\n",
       "      <td>0.012822</td>\n",
       "      <td>-0.088690</td>\n",
       "      <td>0.321527</td>\n",
       "      <td>-0.188313</td>\n",
       "      <td>0.023148</td>\n",
       "      <td>-0.066589</td>\n",
       "      <td>0.068408</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1109.0</th>\n",
       "      <td>-0.012567</td>\n",
       "      <td>0.044235</td>\n",
       "      <td>-0.034262</td>\n",
       "      <td>0.014732</td>\n",
       "      <td>-0.010877</td>\n",
       "      <td>0.006351</td>\n",
       "      <td>-0.022342</td>\n",
       "      <td>0.073960</td>\n",
       "      <td>-0.041287</td>\n",
       "      <td>0.008155</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.019464</td>\n",
       "      <td>0.017875</td>\n",
       "      <td>-0.015584</td>\n",
       "      <td>0.008187</td>\n",
       "      <td>-0.089600</td>\n",
       "      <td>0.306020</td>\n",
       "      <td>-0.186103</td>\n",
       "      <td>0.024226</td>\n",
       "      <td>-0.051694</td>\n",
       "      <td>0.056053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110.0</th>\n",
       "      <td>-0.012638</td>\n",
       "      <td>0.042902</td>\n",
       "      <td>-0.032179</td>\n",
       "      <td>0.016883</td>\n",
       "      <td>-0.014610</td>\n",
       "      <td>0.010634</td>\n",
       "      <td>-0.021678</td>\n",
       "      <td>0.071992</td>\n",
       "      <td>-0.036317</td>\n",
       "      <td>0.013548</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.013538</td>\n",
       "      <td>0.024994</td>\n",
       "      <td>-0.022765</td>\n",
       "      <td>0.011940</td>\n",
       "      <td>-0.096807</td>\n",
       "      <td>0.325933</td>\n",
       "      <td>-0.191077</td>\n",
       "      <td>0.036550</td>\n",
       "      <td>-0.067589</td>\n",
       "      <td>0.062480</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1111.0</th>\n",
       "      <td>-0.014907</td>\n",
       "      <td>0.047401</td>\n",
       "      <td>-0.035259</td>\n",
       "      <td>0.017802</td>\n",
       "      <td>-0.019328</td>\n",
       "      <td>0.012379</td>\n",
       "      <td>-0.025877</td>\n",
       "      <td>0.078515</td>\n",
       "      <td>-0.034360</td>\n",
       "      <td>0.002570</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.007995</td>\n",
       "      <td>0.029181</td>\n",
       "      <td>-0.055187</td>\n",
       "      <td>0.018805</td>\n",
       "      <td>-0.107399</td>\n",
       "      <td>0.339860</td>\n",
       "      <td>-0.194294</td>\n",
       "      <td>0.019429</td>\n",
       "      <td>-0.074855</td>\n",
       "      <td>0.073945</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             value blue_median                                          \\\n",
       "time_step                    3         4         5         6         7   \n",
       "plot_number                                                              \n",
       "1107.0               -0.011184  0.040064 -0.027868  0.011051 -0.009405   \n",
       "1108.0               -0.011775  0.042587 -0.031039  0.013420 -0.010765   \n",
       "1109.0               -0.012567  0.044235 -0.034262  0.014732 -0.010877   \n",
       "1110.0               -0.012638  0.042902 -0.032179  0.016883 -0.014610   \n",
       "1111.0               -0.014907  0.047401 -0.035259  0.017802 -0.019328   \n",
       "\n",
       "                       value green_median                                ...  \\\n",
       "time_step            8                  3         4         5         6  ...   \n",
       "plot_number                                                              ...   \n",
       "1107.0        0.005832          -0.020557  0.071539 -0.036699  0.010031  ...   \n",
       "1108.0        0.003780          -0.020754  0.073087 -0.037642  0.008480  ...   \n",
       "1109.0        0.006351          -0.022342  0.073960 -0.041287  0.008155  ...   \n",
       "1110.0        0.010634          -0.021678  0.071992 -0.036317  0.013548  ...   \n",
       "1111.0        0.012379          -0.025877  0.078515 -0.034360  0.002570  ...   \n",
       "\n",
       "             value red_median                                \\\n",
       "time_step                   5         6         7         8   \n",
       "plot_number                                                   \n",
       "1107.0              -0.015342  0.023259 -0.032606  0.020785   \n",
       "1108.0              -0.013807  0.025891 -0.033949  0.012822   \n",
       "1109.0              -0.019464  0.017875 -0.015584  0.008187   \n",
       "1110.0              -0.013538  0.024994 -0.022765  0.011940   \n",
       "1111.0              -0.007995  0.029181 -0.055187  0.018805   \n",
       "\n",
       "             value rededge_median                                          \\\n",
       "time_step                       3         4         5         6         7   \n",
       "plot_number                                                                 \n",
       "1107.0                  -0.083682  0.302318 -0.175006  0.023571 -0.066351   \n",
       "1108.0                  -0.088690  0.321527 -0.188313  0.023148 -0.066589   \n",
       "1109.0                  -0.089600  0.306020 -0.186103  0.024226 -0.051694   \n",
       "1110.0                  -0.096807  0.325933 -0.191077  0.036550 -0.067589   \n",
       "1111.0                  -0.107399  0.339860 -0.194294  0.019429 -0.074855   \n",
       "\n",
       "                        \n",
       "time_step            8  \n",
       "plot_number             \n",
       "1107.0        0.073858  \n",
       "1108.0        0.068408  \n",
       "1109.0        0.056053  \n",
       "1110.0        0.062480  \n",
       "1111.0        0.073945  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 236,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21pf = df_21diff.pivot_table(index='plot_number ', columns='time_step ', values=['value blue_median','value green_median','value red_median','value rededge_median','value nir_median'], aggfunc='mean')\n",
    "\n",
    "columns_to_remove = [0,1,8,9,10,11,18,19,20,21,28,29,30,31,38,39,40,41,48,49]\n",
    "\n",
    "# Remove first and last observation from each plot\n",
    "df_21p = df_21pf.iloc[:, [i for i in range(df_20pf.shape[1]) if i not in columns_to_remove]]\n",
    "df_21p.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "id": "9c99cb7a-e9ef-4912-927f-4756178e10b0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"6\" halign=\"left\">value blue_median</th>\n",
       "      <th colspan=\"4\" halign=\"left\">value green_median</th>\n",
       "      <th>...</th>\n",
       "      <th colspan=\"4\" halign=\"left\">value red_median</th>\n",
       "      <th colspan=\"6\" halign=\"left\">value rededge_median</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>time_step</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>...</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>plot_number</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>0.001350</td>\n",
       "      <td>-0.000892</td>\n",
       "      <td>-0.002960</td>\n",
       "      <td>0.000103</td>\n",
       "      <td>-0.000196</td>\n",
       "      <td>-0.000763</td>\n",
       "      <td>0.001199</td>\n",
       "      <td>-0.002318</td>\n",
       "      <td>0.004056</td>\n",
       "      <td>-0.010730</td>\n",
       "      <td>...</td>\n",
       "      <td>0.004786</td>\n",
       "      <td>-0.007648</td>\n",
       "      <td>0.007071</td>\n",
       "      <td>0.020039</td>\n",
       "      <td>-0.000059</td>\n",
       "      <td>-0.007422</td>\n",
       "      <td>-0.017835</td>\n",
       "      <td>-0.011743</td>\n",
       "      <td>0.020138</td>\n",
       "      <td>0.001182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1109.0</th>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.000352</td>\n",
       "      <td>-0.002980</td>\n",
       "      <td>-0.000396</td>\n",
       "      <td>-0.000139</td>\n",
       "      <td>-0.000780</td>\n",
       "      <td>0.000443</td>\n",
       "      <td>0.000056</td>\n",
       "      <td>0.004026</td>\n",
       "      <td>-0.013381</td>\n",
       "      <td>...</td>\n",
       "      <td>0.003013</td>\n",
       "      <td>-0.007903</td>\n",
       "      <td>0.009663</td>\n",
       "      <td>0.023169</td>\n",
       "      <td>0.000319</td>\n",
       "      <td>-0.000919</td>\n",
       "      <td>-0.019338</td>\n",
       "      <td>-0.015753</td>\n",
       "      <td>0.026908</td>\n",
       "      <td>-0.005639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110.0</th>\n",
       "      <td>0.000799</td>\n",
       "      <td>-0.000007</td>\n",
       "      <td>-0.002646</td>\n",
       "      <td>-0.000736</td>\n",
       "      <td>0.001885</td>\n",
       "      <td>-0.004562</td>\n",
       "      <td>-0.000061</td>\n",
       "      <td>-0.000636</td>\n",
       "      <td>0.006488</td>\n",
       "      <td>-0.016104</td>\n",
       "      <td>...</td>\n",
       "      <td>0.005734</td>\n",
       "      <td>-0.008973</td>\n",
       "      <td>0.017552</td>\n",
       "      <td>0.008013</td>\n",
       "      <td>-0.001730</td>\n",
       "      <td>-0.009719</td>\n",
       "      <td>-0.009665</td>\n",
       "      <td>-0.022229</td>\n",
       "      <td>0.037995</td>\n",
       "      <td>-0.030505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1112.0</th>\n",
       "      <td>0.001310</td>\n",
       "      <td>0.003206</td>\n",
       "      <td>-0.006159</td>\n",
       "      <td>0.001613</td>\n",
       "      <td>-0.000314</td>\n",
       "      <td>0.000234</td>\n",
       "      <td>0.000564</td>\n",
       "      <td>0.005616</td>\n",
       "      <td>-0.001202</td>\n",
       "      <td>-0.010887</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000555</td>\n",
       "      <td>-0.006273</td>\n",
       "      <td>0.010919</td>\n",
       "      <td>0.013171</td>\n",
       "      <td>-0.002268</td>\n",
       "      <td>0.007768</td>\n",
       "      <td>-0.025873</td>\n",
       "      <td>-0.012764</td>\n",
       "      <td>0.030116</td>\n",
       "      <td>0.000095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1113.0</th>\n",
       "      <td>0.002698</td>\n",
       "      <td>0.005205</td>\n",
       "      <td>-0.005482</td>\n",
       "      <td>-0.001798</td>\n",
       "      <td>0.000925</td>\n",
       "      <td>-0.002631</td>\n",
       "      <td>0.003456</td>\n",
       "      <td>0.009353</td>\n",
       "      <td>0.000840</td>\n",
       "      <td>-0.017267</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001392</td>\n",
       "      <td>-0.010455</td>\n",
       "      <td>0.015915</td>\n",
       "      <td>0.012402</td>\n",
       "      <td>0.006318</td>\n",
       "      <td>0.015118</td>\n",
       "      <td>-0.021724</td>\n",
       "      <td>-0.028973</td>\n",
       "      <td>0.029390</td>\n",
       "      <td>-0.013865</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             value blue_median                                          \\\n",
       "time_step                    3         4         5         6         7   \n",
       "plot_number                                                              \n",
       "1108.0                0.001350 -0.000892 -0.002960  0.000103 -0.000196   \n",
       "1109.0                0.001238  0.000352 -0.002980 -0.000396 -0.000139   \n",
       "1110.0                0.000799 -0.000007 -0.002646 -0.000736  0.001885   \n",
       "1112.0                0.001310  0.003206 -0.006159  0.001613 -0.000314   \n",
       "1113.0                0.002698  0.005205 -0.005482 -0.001798  0.000925   \n",
       "\n",
       "                       value green_median                                ...  \\\n",
       "time_step            8                  3         4         5         6  ...   \n",
       "plot_number                                                              ...   \n",
       "1108.0       -0.000763           0.001199 -0.002318  0.004056 -0.010730  ...   \n",
       "1109.0       -0.000780           0.000443  0.000056  0.004026 -0.013381  ...   \n",
       "1110.0       -0.004562          -0.000061 -0.000636  0.006488 -0.016104  ...   \n",
       "1112.0        0.000234           0.000564  0.005616 -0.001202 -0.010887  ...   \n",
       "1113.0       -0.002631           0.003456  0.009353  0.000840 -0.017267  ...   \n",
       "\n",
       "             value red_median                                \\\n",
       "time_step                   5         6         7         8   \n",
       "plot_number                                                   \n",
       "1108.0               0.004786 -0.007648  0.007071  0.020039   \n",
       "1109.0               0.003013 -0.007903  0.009663  0.023169   \n",
       "1110.0               0.005734 -0.008973  0.017552  0.008013   \n",
       "1112.0               0.000555 -0.006273  0.010919  0.013171   \n",
       "1113.0               0.001392 -0.010455  0.015915  0.012402   \n",
       "\n",
       "             value rededge_median                                          \\\n",
       "time_step                       3         4         5         6         7   \n",
       "plot_number                                                                 \n",
       "1108.0                  -0.000059 -0.007422 -0.017835 -0.011743  0.020138   \n",
       "1109.0                   0.000319 -0.000919 -0.019338 -0.015753  0.026908   \n",
       "1110.0                  -0.001730 -0.009719 -0.009665 -0.022229  0.037995   \n",
       "1112.0                  -0.002268  0.007768 -0.025873 -0.012764  0.030116   \n",
       "1113.0                   0.006318  0.015118 -0.021724 -0.028973  0.029390   \n",
       "\n",
       "                        \n",
       "time_step            8  \n",
       "plot_number             \n",
       "1108.0        0.001182  \n",
       "1109.0       -0.005639  \n",
       "1110.0       -0.030505  \n",
       "1112.0        0.000095  \n",
       "1113.0       -0.013865  \n",
       "\n",
       "[5 rows x 30 columns]"
      ]
     },
     "execution_count": 237,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_22pf = df_22diff.pivot_table(index='plot_number ', columns='time_step ', values=['value blue_median','value green_median','value red_median','value rededge_median','value nir_median'], aggfunc='mean')\n",
    "\n",
    "columns_to_remove = [0,1,8,9,10,11,18,19,20,21,28,29,30,31,38,39,40,41,48,49]\n",
    "\n",
    "# Remove first and last observation from each plot\n",
    "df_22p = df_22pf.iloc[:, [i for i in range(df_22pf.shape[1]) if i not in columns_to_remove]]\n",
    "df_22p.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f8eaa177-2c19-4227-820e-86a1a5bde7aa",
   "metadata": {},
   "source": [
    "## Concatenating the Target Variable (y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "id": "51cf92ac-9f70-468d-b7ae-2fe85986e210",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_y_raw = pd.read_excel('masbasis_agro_complete.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "id": "bacdbd96-b934-401c-9266-f041ca5bad0e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>env</th>\n",
       "      <th>season</th>\n",
       "      <th>location</th>\n",
       "      <th>plot_number</th>\n",
       "      <th>plot_type</th>\n",
       "      <th>line_number</th>\n",
       "      <th>masbasis2015</th>\n",
       "      <th>rep</th>\n",
       "      <th>block</th>\n",
       "      <th>column</th>\n",
       "      <th>DH_dss</th>\n",
       "      <th>DM_dss</th>\n",
       "      <th>GY_g_m2</th>\n",
       "      <th>PH_cm</th>\n",
       "      <th>GPC_pct</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3801</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1105</td>\n",
       "      <td>yield</td>\n",
       "      <td>6.0</td>\n",
       "      <td>1029.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>713.333333</td>\n",
       "      <td>81.67</td>\n",
       "      <td>10.4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3802</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1106</td>\n",
       "      <td>yield</td>\n",
       "      <td>1543.0</td>\n",
       "      <td>1543.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>119.0</td>\n",
       "      <td>677.333333</td>\n",
       "      <td>86.67</td>\n",
       "      <td>12.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3803</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1107</td>\n",
       "      <td>yield</td>\n",
       "      <td>1338.0</td>\n",
       "      <td>1338.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>120.0</td>\n",
       "      <td>361.333333</td>\n",
       "      <td>95.00</td>\n",
       "      <td>15.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3804</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1108</td>\n",
       "      <td>yield</td>\n",
       "      <td>1526.0</td>\n",
       "      <td>1526.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>67.0</td>\n",
       "      <td>117.0</td>\n",
       "      <td>697.333333</td>\n",
       "      <td>86.67</td>\n",
       "      <td>10.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3805</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1109</td>\n",
       "      <td>yield</td>\n",
       "      <td>1614.0</td>\n",
       "      <td>1614.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>664.000000</td>\n",
       "      <td>76.67</td>\n",
       "      <td>11.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4372</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1878</td>\n",
       "      <td>yield</td>\n",
       "      <td>1504.0</td>\n",
       "      <td>1504.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>78.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>116.0</td>\n",
       "      <td>646.666667</td>\n",
       "      <td>78.33</td>\n",
       "      <td>11.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4373</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1879</td>\n",
       "      <td>yield</td>\n",
       "      <td>1330.0</td>\n",
       "      <td>1330.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>79.0</td>\n",
       "      <td>64.0</td>\n",
       "      <td>119.0</td>\n",
       "      <td>478.666667</td>\n",
       "      <td>85.00</td>\n",
       "      <td>13.8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4374</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1880</td>\n",
       "      <td>yield</td>\n",
       "      <td>1406.0</td>\n",
       "      <td>1406.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>80.0</td>\n",
       "      <td>68.0</td>\n",
       "      <td>117.0</td>\n",
       "      <td>732.000000</td>\n",
       "      <td>83.33</td>\n",
       "      <td>10.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4375</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1881</td>\n",
       "      <td>yield</td>\n",
       "      <td>1308.0</td>\n",
       "      <td>1308.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>81.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>116.0</td>\n",
       "      <td>600.000000</td>\n",
       "      <td>88.33</td>\n",
       "      <td>12.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4376</th>\n",
       "      <td>2020_vollebekk</td>\n",
       "      <td>2020</td>\n",
       "      <td>vollebekk</td>\n",
       "      <td>1882</td>\n",
       "      <td>yield</td>\n",
       "      <td>1532.0</td>\n",
       "      <td>1532.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>82.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>115.0</td>\n",
       "      <td>714.666667</td>\n",
       "      <td>90.00</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>576 rows × 15 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 env  season   location  plot_number plot_type  line_number  \\\n",
       "3801  2020_vollebekk    2020  vollebekk         1105     yield          6.0   \n",
       "3802  2020_vollebekk    2020  vollebekk         1106     yield       1543.0   \n",
       "3803  2020_vollebekk    2020  vollebekk         1107     yield       1338.0   \n",
       "3804  2020_vollebekk    2020  vollebekk         1108     yield       1526.0   \n",
       "3805  2020_vollebekk    2020  vollebekk         1109     yield       1614.0   \n",
       "...              ...     ...        ...          ...       ...          ...   \n",
       "4372  2020_vollebekk    2020  vollebekk         1878     yield       1504.0   \n",
       "4373  2020_vollebekk    2020  vollebekk         1879     yield       1330.0   \n",
       "4374  2020_vollebekk    2020  vollebekk         1880     yield       1406.0   \n",
       "4375  2020_vollebekk    2020  vollebekk         1881     yield       1308.0   \n",
       "4376  2020_vollebekk    2020  vollebekk         1882     yield       1532.0   \n",
       "\n",
       "      masbasis2015  rep  block  column  DH_dss  DM_dss     GY_g_m2  PH_cm  \\\n",
       "3801        1029.0  1.0    1.0     5.0    66.0   118.0  713.333333  81.67   \n",
       "3802        1543.0  1.0    1.0     6.0    68.0   119.0  677.333333  86.67   \n",
       "3803        1338.0  1.0    1.0     7.0    66.0   120.0  361.333333  95.00   \n",
       "3804        1526.0  1.0    1.0     8.0    67.0   117.0  697.333333  86.67   \n",
       "3805        1614.0  1.0    1.0     9.0    68.0   118.0  664.000000  76.67   \n",
       "...            ...  ...    ...     ...     ...     ...         ...    ...   \n",
       "4372        1504.0  2.0   48.0    78.0    66.0   116.0  646.666667  78.33   \n",
       "4373        1330.0  2.0   48.0    79.0    64.0   119.0  478.666667  85.00   \n",
       "4374        1406.0  2.0   48.0    80.0    68.0   117.0  732.000000  83.33   \n",
       "4375        1308.0  2.0   48.0    81.0    65.0   116.0  600.000000  88.33   \n",
       "4376        1532.0  2.0   48.0    82.0    66.0   115.0  714.666667  90.00   \n",
       "\n",
       "      GPC_pct  \n",
       "3801     10.4  \n",
       "3802     12.1  \n",
       "3803     15.5  \n",
       "3804     10.8  \n",
       "3805     11.5  \n",
       "...       ...  \n",
       "4372     11.3  \n",
       "4373     13.8  \n",
       "4374     10.3  \n",
       "4375     12.1  \n",
       "4376     12.0  \n",
       "\n",
       "[576 rows x 15 columns]"
      ]
     },
     "execution_count": 239,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_y_raw.loc[(df_y_raw['season']==2020) & (df_y_raw['location']=='vollebekk') & (df_y_raw['plot_type']=='yield')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "id": "f5dee21c-8e4d-46e2-9b1a-32878ec0f64c",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_y20 = df_y_raw.loc[(df_y_raw['season']==2020) & (df_y_raw['location']=='vollebekk') & (df_y_raw['plot_type']=='yield')]\n",
    "df_y20 = df_y20[['plot_number','GY_g_m2']]\n",
    "df_y20 = df_y20.set_index('plot_number')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "id": "5d8bc1fb-244a-4326-8667-9acaf856a6f1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 241,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_y20.isna().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "id": "bf7b727c-621f-4f13-ae7e-c2420db6ef49",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(422, 30)"
      ]
     },
     "execution_count": 242,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_20p.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "id": "9a132706-1ab0-4403-ab90-7b8bc4c8a2ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_conc20 = pd.concat([df_20p, df_y20], axis='columns')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "id": "f0aa79ce-0ae4-4dd7-bf37-689bbc228ee4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(419, 31)"
      ]
     },
     "execution_count": 244,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_conc20 = df_conc20.dropna()\n",
    "df_conc20.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "id": "512ef76e-2e12-47d6-9cc7-7c5c859ec781",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_y21 = df_y_raw.loc[(df_y_raw['season']==2021) & (df_y_raw['location']=='vollebekk') & (df_y_raw['plot_type']=='yield')]\n",
    "df_y21 = df_y21[['plot_number','GY_g_m2']]\n",
    "df_y21 = df_y21.set_index('plot_number')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "id": "830f6ea1-6cc8-4035-8ad2-c66801bf014d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(576, 1)"
      ]
     },
     "execution_count": 246,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_y21.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "id": "db9991df-22a3-430b-bf75-639d7c70047c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10"
      ]
     },
     "execution_count": 247,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_y21.isna().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "id": "4d677166-bfb6-4958-ac69-9792226d0fd0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(508, 30)"
      ]
     },
     "execution_count": 248,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_21p.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "id": "0af923c7-04aa-4271-83fd-5359a9f99bc5",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_conc21 = pd.concat([df_21p, df_y21], axis='columns')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 250,
   "id": "823fd03a-805a-45da-9aa0-bd32bfb2cae9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(501, 31)"
      ]
     },
     "execution_count": 250,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_conc21 = df_conc21.dropna()\n",
    "df_conc21.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "id": "2ae54f88-c6c7-4ee0-8e02-8d67880d8cd1",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_y22 = df_y_raw.loc[(df_y_raw['season']==2022) & (df_y_raw['location']=='vollebekk') & (df_y_raw['plot_type']=='yield')]\n",
    "df_y22 = df_y22[['plot_number','GY_g_m2']]\n",
    "df_y22 = df_y22.set_index('plot_number')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "id": "6dccd57c-4cc6-47bc-970a-86cb47f4cf16",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(576, 1)"
      ]
     },
     "execution_count": 252,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_y22.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "id": "c77bf60e-67d4-4613-ba0b-b5e48a077a7b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 253,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_y22.isna().sum().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "id": "9e6b0aac-ab3f-4a0b-91f1-8d4fa1be411b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(513, 30)"
      ]
     },
     "execution_count": 254,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_22p.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "id": "c463267f-8543-4414-90f7-3cd36e3883e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_conc22 = pd.concat([df_22p, df_y22], axis='columns')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "id": "ef7047e9-fa45-4754-bf39-1614d5c1d263",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(512, 31)"
      ]
     },
     "execution_count": 256,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_conc22 = df_conc22.dropna()\n",
    "df_conc22.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aa2c2618-c8cb-4834-b495-d4b1bca1dd79",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 257,
   "id": "5384c4ee-aaeb-4639-ba66-fe610279ef34",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_conc = pd.concat([df_conc20, df_conc21, df_conc22], axis='rows')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "id": "3b5f4178-bf22-4ca2-84d7-9ffb4e7cc711",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1432, 31)"
      ]
     },
     "execution_count": 258,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_conc.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "id": "11737cbe-9442-49eb-bdbc-29ab14665c60",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>(value blue_median, 3)</th>\n",
       "      <th>(value blue_median, 4)</th>\n",
       "      <th>(value blue_median, 5)</th>\n",
       "      <th>(value blue_median, 6)</th>\n",
       "      <th>(value blue_median, 7)</th>\n",
       "      <th>(value blue_median, 8)</th>\n",
       "      <th>(value green_median, 3)</th>\n",
       "      <th>(value green_median, 4)</th>\n",
       "      <th>(value green_median, 5)</th>\n",
       "      <th>(value green_median, 6)</th>\n",
       "      <th>...</th>\n",
       "      <th>(value red_median, 6)</th>\n",
       "      <th>(value red_median, 7)</th>\n",
       "      <th>(value red_median, 8)</th>\n",
       "      <th>(value rededge_median, 3)</th>\n",
       "      <th>(value rededge_median, 4)</th>\n",
       "      <th>(value rededge_median, 5)</th>\n",
       "      <th>(value rededge_median, 6)</th>\n",
       "      <th>(value rededge_median, 7)</th>\n",
       "      <th>(value rededge_median, 8)</th>\n",
       "      <th>GY_g_m2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1105.0</th>\n",
       "      <td>-0.008605</td>\n",
       "      <td>-0.009388</td>\n",
       "      <td>0.022602</td>\n",
       "      <td>-0.004606</td>\n",
       "      <td>-0.008749</td>\n",
       "      <td>-0.000972</td>\n",
       "      <td>-0.018040</td>\n",
       "      <td>-0.011211</td>\n",
       "      <td>0.046606</td>\n",
       "      <td>-0.011864</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.013663</td>\n",
       "      <td>-0.014164</td>\n",
       "      <td>0.004590</td>\n",
       "      <td>-0.040992</td>\n",
       "      <td>-0.027322</td>\n",
       "      <td>0.111806</td>\n",
       "      <td>-0.028127</td>\n",
       "      <td>-0.043465</td>\n",
       "      <td>-0.008273</td>\n",
       "      <td>713.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1106.0</th>\n",
       "      <td>-0.006848</td>\n",
       "      <td>-0.007819</td>\n",
       "      <td>0.023717</td>\n",
       "      <td>-0.009342</td>\n",
       "      <td>-0.008228</td>\n",
       "      <td>0.004907</td>\n",
       "      <td>-0.015791</td>\n",
       "      <td>-0.007327</td>\n",
       "      <td>0.049493</td>\n",
       "      <td>-0.021909</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.017280</td>\n",
       "      <td>-0.008371</td>\n",
       "      <td>0.017430</td>\n",
       "      <td>-0.041287</td>\n",
       "      <td>-0.023628</td>\n",
       "      <td>0.114447</td>\n",
       "      <td>-0.039549</td>\n",
       "      <td>-0.033043</td>\n",
       "      <td>0.019659</td>\n",
       "      <td>677.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1108.0</th>\n",
       "      <td>-0.005938</td>\n",
       "      <td>-0.002933</td>\n",
       "      <td>0.020841</td>\n",
       "      <td>-0.007454</td>\n",
       "      <td>-0.009837</td>\n",
       "      <td>-0.000877</td>\n",
       "      <td>-0.015265</td>\n",
       "      <td>-0.004520</td>\n",
       "      <td>0.058495</td>\n",
       "      <td>-0.026264</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.020196</td>\n",
       "      <td>-0.017759</td>\n",
       "      <td>0.008623</td>\n",
       "      <td>-0.039108</td>\n",
       "      <td>-0.023412</td>\n",
       "      <td>0.134376</td>\n",
       "      <td>-0.045555</td>\n",
       "      <td>-0.055735</td>\n",
       "      <td>-0.004461</td>\n",
       "      <td>697.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1109.0</th>\n",
       "      <td>-0.006748</td>\n",
       "      <td>-0.003985</td>\n",
       "      <td>0.023220</td>\n",
       "      <td>-0.009352</td>\n",
       "      <td>-0.012201</td>\n",
       "      <td>-0.001166</td>\n",
       "      <td>-0.017631</td>\n",
       "      <td>-0.007990</td>\n",
       "      <td>0.056392</td>\n",
       "      <td>-0.023796</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.011797</td>\n",
       "      <td>-0.018655</td>\n",
       "      <td>-0.002589</td>\n",
       "      <td>-0.045216</td>\n",
       "      <td>-0.032807</td>\n",
       "      <td>0.127332</td>\n",
       "      <td>-0.035732</td>\n",
       "      <td>-0.053243</td>\n",
       "      <td>-0.010882</td>\n",
       "      <td>664.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1110.0</th>\n",
       "      <td>-0.003405</td>\n",
       "      <td>-0.003346</td>\n",
       "      <td>0.022287</td>\n",
       "      <td>-0.012247</td>\n",
       "      <td>-0.010869</td>\n",
       "      <td>0.002972</td>\n",
       "      <td>-0.013574</td>\n",
       "      <td>-0.005082</td>\n",
       "      <td>0.051582</td>\n",
       "      <td>-0.016651</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.002846</td>\n",
       "      <td>-0.010452</td>\n",
       "      <td>-0.003176</td>\n",
       "      <td>-0.037416</td>\n",
       "      <td>-0.024271</td>\n",
       "      <td>0.116596</td>\n",
       "      <td>-0.018274</td>\n",
       "      <td>-0.049061</td>\n",
       "      <td>-0.029272</td>\n",
       "      <td>456.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1882.0</th>\n",
       "      <td>0.001381</td>\n",
       "      <td>-0.000350</td>\n",
       "      <td>-0.004224</td>\n",
       "      <td>0.001462</td>\n",
       "      <td>0.001411</td>\n",
       "      <td>-0.000426</td>\n",
       "      <td>0.002740</td>\n",
       "      <td>-0.000904</td>\n",
       "      <td>-0.001475</td>\n",
       "      <td>-0.007327</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.003923</td>\n",
       "      <td>0.012880</td>\n",
       "      <td>0.019156</td>\n",
       "      <td>0.008520</td>\n",
       "      <td>-0.008767</td>\n",
       "      <td>-0.025223</td>\n",
       "      <td>0.001088</td>\n",
       "      <td>0.026121</td>\n",
       "      <td>0.007800</td>\n",
       "      <td>652.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1883.0</th>\n",
       "      <td>0.000581</td>\n",
       "      <td>-0.001471</td>\n",
       "      <td>-0.002261</td>\n",
       "      <td>0.001367</td>\n",
       "      <td>-0.001142</td>\n",
       "      <td>0.001683</td>\n",
       "      <td>0.001059</td>\n",
       "      <td>-0.003524</td>\n",
       "      <td>0.003763</td>\n",
       "      <td>-0.010374</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.005594</td>\n",
       "      <td>0.008898</td>\n",
       "      <td>0.022345</td>\n",
       "      <td>0.000720</td>\n",
       "      <td>-0.019362</td>\n",
       "      <td>-0.009192</td>\n",
       "      <td>-0.009811</td>\n",
       "      <td>0.024856</td>\n",
       "      <td>0.010733</td>\n",
       "      <td>612.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1884.0</th>\n",
       "      <td>0.001367</td>\n",
       "      <td>0.000379</td>\n",
       "      <td>-0.004096</td>\n",
       "      <td>0.001249</td>\n",
       "      <td>-0.002513</td>\n",
       "      <td>0.002677</td>\n",
       "      <td>0.002595</td>\n",
       "      <td>0.000939</td>\n",
       "      <td>-0.001702</td>\n",
       "      <td>-0.008704</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.004131</td>\n",
       "      <td>0.002833</td>\n",
       "      <td>0.011902</td>\n",
       "      <td>0.005203</td>\n",
       "      <td>-0.010704</td>\n",
       "      <td>-0.023123</td>\n",
       "      <td>-0.003949</td>\n",
       "      <td>0.011180</td>\n",
       "      <td>0.026749</td>\n",
       "      <td>716.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1885.0</th>\n",
       "      <td>0.001271</td>\n",
       "      <td>0.000425</td>\n",
       "      <td>-0.003215</td>\n",
       "      <td>0.000721</td>\n",
       "      <td>-0.000067</td>\n",
       "      <td>0.003488</td>\n",
       "      <td>0.001750</td>\n",
       "      <td>0.000743</td>\n",
       "      <td>0.001015</td>\n",
       "      <td>-0.010625</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.005025</td>\n",
       "      <td>0.008243</td>\n",
       "      <td>0.020460</td>\n",
       "      <td>0.003093</td>\n",
       "      <td>-0.012047</td>\n",
       "      <td>-0.016640</td>\n",
       "      <td>-0.008813</td>\n",
       "      <td>0.022339</td>\n",
       "      <td>0.021138</td>\n",
       "      <td>681.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1886.0</th>\n",
       "      <td>-0.000999</td>\n",
       "      <td>-0.000644</td>\n",
       "      <td>-0.002092</td>\n",
       "      <td>0.001838</td>\n",
       "      <td>-0.001035</td>\n",
       "      <td>0.001617</td>\n",
       "      <td>-0.001954</td>\n",
       "      <td>-0.001922</td>\n",
       "      <td>0.001027</td>\n",
       "      <td>-0.005793</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.002380</td>\n",
       "      <td>0.006875</td>\n",
       "      <td>0.005048</td>\n",
       "      <td>-0.010104</td>\n",
       "      <td>-0.016190</td>\n",
       "      <td>-0.015816</td>\n",
       "      <td>0.002717</td>\n",
       "      <td>0.019254</td>\n",
       "      <td>0.008961</td>\n",
       "      <td>589.333333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1432 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        (value blue_median, 3)  (value blue_median, 4)  \\\n",
       "1105.0               -0.008605               -0.009388   \n",
       "1106.0               -0.006848               -0.007819   \n",
       "1108.0               -0.005938               -0.002933   \n",
       "1109.0               -0.006748               -0.003985   \n",
       "1110.0               -0.003405               -0.003346   \n",
       "...                        ...                     ...   \n",
       "1882.0                0.001381               -0.000350   \n",
       "1883.0                0.000581               -0.001471   \n",
       "1884.0                0.001367                0.000379   \n",
       "1885.0                0.001271                0.000425   \n",
       "1886.0               -0.000999               -0.000644   \n",
       "\n",
       "        (value blue_median, 5)  (value blue_median, 6)  \\\n",
       "1105.0                0.022602               -0.004606   \n",
       "1106.0                0.023717               -0.009342   \n",
       "1108.0                0.020841               -0.007454   \n",
       "1109.0                0.023220               -0.009352   \n",
       "1110.0                0.022287               -0.012247   \n",
       "...                        ...                     ...   \n",
       "1882.0               -0.004224                0.001462   \n",
       "1883.0               -0.002261                0.001367   \n",
       "1884.0               -0.004096                0.001249   \n",
       "1885.0               -0.003215                0.000721   \n",
       "1886.0               -0.002092                0.001838   \n",
       "\n",
       "        (value blue_median, 7)  (value blue_median, 8)  \\\n",
       "1105.0               -0.008749               -0.000972   \n",
       "1106.0               -0.008228                0.004907   \n",
       "1108.0               -0.009837               -0.000877   \n",
       "1109.0               -0.012201               -0.001166   \n",
       "1110.0               -0.010869                0.002972   \n",
       "...                        ...                     ...   \n",
       "1882.0                0.001411               -0.000426   \n",
       "1883.0               -0.001142                0.001683   \n",
       "1884.0               -0.002513                0.002677   \n",
       "1885.0               -0.000067                0.003488   \n",
       "1886.0               -0.001035                0.001617   \n",
       "\n",
       "        (value green_median, 3)  (value green_median, 4)  \\\n",
       "1105.0                -0.018040                -0.011211   \n",
       "1106.0                -0.015791                -0.007327   \n",
       "1108.0                -0.015265                -0.004520   \n",
       "1109.0                -0.017631                -0.007990   \n",
       "1110.0                -0.013574                -0.005082   \n",
       "...                         ...                      ...   \n",
       "1882.0                 0.002740                -0.000904   \n",
       "1883.0                 0.001059                -0.003524   \n",
       "1884.0                 0.002595                 0.000939   \n",
       "1885.0                 0.001750                 0.000743   \n",
       "1886.0                -0.001954                -0.001922   \n",
       "\n",
       "        (value green_median, 5)  (value green_median, 6)  ...  \\\n",
       "1105.0                 0.046606                -0.011864  ...   \n",
       "1106.0                 0.049493                -0.021909  ...   \n",
       "1108.0                 0.058495                -0.026264  ...   \n",
       "1109.0                 0.056392                -0.023796  ...   \n",
       "1110.0                 0.051582                -0.016651  ...   \n",
       "...                         ...                      ...  ...   \n",
       "1882.0                -0.001475                -0.007327  ...   \n",
       "1883.0                 0.003763                -0.010374  ...   \n",
       "1884.0                -0.001702                -0.008704  ...   \n",
       "1885.0                 0.001015                -0.010625  ...   \n",
       "1886.0                 0.001027                -0.005793  ...   \n",
       "\n",
       "        (value red_median, 6)  (value red_median, 7)  (value red_median, 8)  \\\n",
       "1105.0              -0.013663              -0.014164               0.004590   \n",
       "1106.0              -0.017280              -0.008371               0.017430   \n",
       "1108.0              -0.020196              -0.017759               0.008623   \n",
       "1109.0              -0.011797              -0.018655              -0.002589   \n",
       "1110.0              -0.002846              -0.010452              -0.003176   \n",
       "...                       ...                    ...                    ...   \n",
       "1882.0              -0.003923               0.012880               0.019156   \n",
       "1883.0              -0.005594               0.008898               0.022345   \n",
       "1884.0              -0.004131               0.002833               0.011902   \n",
       "1885.0              -0.005025               0.008243               0.020460   \n",
       "1886.0              -0.002380               0.006875               0.005048   \n",
       "\n",
       "        (value rededge_median, 3)  (value rededge_median, 4)  \\\n",
       "1105.0                  -0.040992                  -0.027322   \n",
       "1106.0                  -0.041287                  -0.023628   \n",
       "1108.0                  -0.039108                  -0.023412   \n",
       "1109.0                  -0.045216                  -0.032807   \n",
       "1110.0                  -0.037416                  -0.024271   \n",
       "...                           ...                        ...   \n",
       "1882.0                   0.008520                  -0.008767   \n",
       "1883.0                   0.000720                  -0.019362   \n",
       "1884.0                   0.005203                  -0.010704   \n",
       "1885.0                   0.003093                  -0.012047   \n",
       "1886.0                  -0.010104                  -0.016190   \n",
       "\n",
       "        (value rededge_median, 5)  (value rededge_median, 6)  \\\n",
       "1105.0                   0.111806                  -0.028127   \n",
       "1106.0                   0.114447                  -0.039549   \n",
       "1108.0                   0.134376                  -0.045555   \n",
       "1109.0                   0.127332                  -0.035732   \n",
       "1110.0                   0.116596                  -0.018274   \n",
       "...                           ...                        ...   \n",
       "1882.0                  -0.025223                   0.001088   \n",
       "1883.0                  -0.009192                  -0.009811   \n",
       "1884.0                  -0.023123                  -0.003949   \n",
       "1885.0                  -0.016640                  -0.008813   \n",
       "1886.0                  -0.015816                   0.002717   \n",
       "\n",
       "        (value rededge_median, 7)  (value rededge_median, 8)     GY_g_m2  \n",
       "1105.0                  -0.043465                  -0.008273  713.333333  \n",
       "1106.0                  -0.033043                   0.019659  677.333333  \n",
       "1108.0                  -0.055735                  -0.004461  697.333333  \n",
       "1109.0                  -0.053243                  -0.010882  664.000000  \n",
       "1110.0                  -0.049061                  -0.029272  456.000000  \n",
       "...                           ...                        ...         ...  \n",
       "1882.0                   0.026121                   0.007800  652.000000  \n",
       "1883.0                   0.024856                   0.010733  612.000000  \n",
       "1884.0                   0.011180                   0.026749  716.000000  \n",
       "1885.0                   0.022339                   0.021138  681.333333  \n",
       "1886.0                   0.019254                   0.008961  589.333333  \n",
       "\n",
       "[1432 rows x 31 columns]"
      ]
     },
     "execution_count": 259,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_conc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "id": "719ad95c-c4b3-4103-a6d9-406f13d2e2fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "#df_conc.to_excel('df_concatenated.xlsx')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d5819134-610c-4ba7-b26c-e33f4ed904dd",
   "metadata": {},
   "source": [
    "## X and Y separation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "id": "8c18fe94-b260-4aab-8418-a8fbe662ab6c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1432, 30)"
      ]
     },
     "execution_count": 261,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = df_conc.iloc[:,0:30].values\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 262,
   "id": "c0a3b00e-9592-46c4-aae8-682f2aa59540",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1432,)"
      ]
     },
     "execution_count": 262,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = df_conc.iloc[:,30].values\n",
    "y.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46d1ad16-6048-458a-8666-30546c461efa",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Standardization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "id": "38f488e1-08d1-4a4d-859b-3f67894fb434",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1432, 30)"
      ]
     },
     "execution_count": 264,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc = StandardScaler()\n",
    "\n",
    "X_sc = sc.fit_transform(X)\n",
    "\n",
    "#X_sc = X\n",
    "X_sc.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "69fe5490-671f-4341-a2e9-e56d357dda8f",
   "metadata": {},
   "source": [
    "## Train Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 265,
   "id": "e7e4c1f4-9b8d-4cce-af71-1430de9a1597",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1145, 30) (287, 30) (1145,) (287,)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train_a, X_test, y_train_a, y_test = train_test_split(X_sc, y, test_size = 0.2, random_state = 1)\n",
    "\n",
    "print(X_train_a.shape, X_test.shape, y_train_a.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "id": "bba9a409-3ec9-47b0-9c70-439931ae3463",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(916, 30) (229, 30) (916,) (229,)\n"
     ]
    }
   ],
   "source": [
    "### Validation Split ###\n",
    "\n",
    "X_train, X_val, y_train, y_val = train_test_split(X_train_a, y_train_a, test_size=0.2, random_state = 5)\n",
    "print(X_train.shape, X_val.shape, y_train.shape, y_val.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "id": "385044ec-e7af-4849-bbe2-5ba161160ba0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_loss(history):\n",
    "  plt.plot(history.history['loss'], label='loss')\n",
    "  plt.plot(history.history['val_loss'], label='val_loss')\n",
    "  plt.xlabel('Epoch')\n",
    "  plt.ylabel('Error')\n",
    "  plt.legend()\n",
    "  plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "id": "44667537-8def-4278-bf5d-e61a6c3343ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import RepeatedStratifiedKFold\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.feature_selection import RFE\n",
    "from sklearn.ensemble import GradientBoostingRegressor\n",
    "from sklearn.metrics import mean_squared_error\n",
    "\n",
    "model = GradientBoostingRegressor(random_state=0)\n",
    "pipe_control = Pipeline([('Model', model)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 269,
   "id": "c97ba85f-6095-4f26-848d-e40195adb3b5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: total: 562 ms\n",
      "Wall time: 1.04 s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Pipeline(steps=[('Model', GradientBoostingRegressor(random_state=0))])"
      ]
     },
     "execution_count": 269,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "pipe_control.fit(X_train_a, y_train_a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "id": "1eadee3a-357a-40da-af42-0f9ecdfb682b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7994227689368227"
      ]
     },
     "execution_count": 270,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe_control.score(X_train_a, y_train_a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "id": "5151cc81-c74e-4e3f-9b4c-b1019c26fee2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.539801553513102"
      ]
     },
     "execution_count": 271,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pipe_control.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 272,
   "id": "bf26c853-287d-41b6-b24b-732fe6c5f6ea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "44.91716778007961\n"
     ]
    }
   ],
   "source": [
    "y_pred_tr = pipe_control.predict(X_train_a)\n",
    "pipe_rmse_train = np.sqrt(mean_squared_error(y_train_a, y_pred_tr))\n",
    "print(pipe_rmse_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "id": "8bbbaa80-fc4a-4a68-a8a6-8f0a21c6bb0a",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = pipe_control.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "id": "32139775-72ac-4601-96d5-8aab4ef50215",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "66.02444680977898\n"
     ]
    }
   ],
   "source": [
    "pipe_rmse_test = np.sqrt(mean_squared_error(y_test, y_pred))\n",
    "print(pipe_rmse_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "id": "f1762c81-4e25-4fd7-9c87-8636953053af",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsgAAALICAYAAABiqwZ2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAcp0lEQVR4nO3df7Cmd1nf8c9FFgUVNBk2aQakx2rqSK2E6YIRtEUQjcYxwRqLU3WnQ0211gG12lX/sPavOO1YRltlUmRYKmKjkiGyrTSGXzpFQhJRgkFjNdAMmWzEX2ApDuHqH+cOXm7Obs4muc9zTvJ6zZx5nuf7/LpO7j2b99x7n+eu7g4AALDtcZseAAAA9hOBDAAAg0AGAIBBIAMAwCCQAQBgOLTpAXbjKU95Sm9tbW16DAAAHkVuueWWP+7uw6euH4hA3trays0337zpMQAAeBSpqg/stO4QCwAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwHBo0wMAD93WsROrv8edV1+2+nsAwH5iDzIAAAwCGQAABoEMAACDQAYAgEEgAwDAIJABAGAQyAAAMAhkAAAYVj1RSFXdmeQjSe5L8onuPlJV5yX5b0m2ktyZ5Ju7+0/XnAMAAHZrL/Ygf2V3X9zdR5bbx5Lc2N0XJblxuQ0AAPvCJg6xuDzJ8eX68SRXbGAGAADY0dqB3En+Z1XdUlVXLWsXdPfdSbJcnr/yDAAAsGurHoOc5Hnd/aGqOj/JDVX1/t0+cQnqq5Lk6U9/+lrzAQDA37DqHuTu/tByeTLJdUmek+SeqrowSZbLk6d57jXdfaS7jxw+fHjNMQEA4FNWC+Sq+syqetL915N8dZLbklyf5OjysKNJ3rjWDAAAcLbWPMTigiTXVdX97/Pz3f2rVfXuJNdW1UuTfDDJlSvOAAAAZ2W1QO7uP0zyzB3WP5zkhWu9LwAAPBzOpAcAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYDi06QHg0Wzr2IlNjwAAnCV7kAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAYfVArqpzquq3qupNy+3zquqGqrpjuTx37RkAAGC39mIP8suS3D5uH0tyY3dflOTG5TYAAOwLqwZyVT0tyWVJXjWWL09yfLl+PMkVa84AAABnY+09yK9I8oNJPjnWLujuu5NkuTx/pydW1VVVdXNV3XzvvfeuPCYAAGxbLZCr6uuTnOzuWx7K87v7mu4+0t1HDh8+/AhPBwAAOzu04ms/L8k3VNXXJXlCkidX1c8luaeqLuzuu6vqwiQnV5wBAADOymp7kLv7h7r7ad29leQlSd7S3d+a5PokR5eHHU3yxrVmAACAs7WJz0G+OsmLquqOJC9abgMAwL6w5iEWn9Ldb0vytuX6h5O8cC/eFwAAzpYz6QEAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGA5tegDYpK1jJzY9AgCwz9iDDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgOLTpAYD9bevYiVVf/86rL1v19QHgbNmDDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBhtUCuqidU1U1V9dtV9b6q+rFl/byquqGq7lguz11rBgAAOFtr7kH+eJIXdPczk1yc5NKquiTJsSQ3dvdFSW5cbgMAwL6wWiD3to8uNx+/fHWSy5McX9aPJ7lirRkAAOBsrXoMclWdU1XvSXIyyQ3d/a4kF3T33UmyXJ5/mudeVVU3V9XN995775pjAgDAp6wayN19X3dfnORpSZ5TVV98Fs+9pruPdPeRw4cPrzYjAABMe/IpFt39Z0neluTSJPdU1YVJslye3IsZAABgN9b8FIvDVfU5y/UnJvmqJO9Pcn2So8vDjiZ541ozAADA2Tq04mtfmOR4VZ2T7RC/trvfVFXvTHJtVb00yQeTXLniDAAAcFZWC+Tu/p0kz9ph/cNJXrjW+wIAwMPhTHoAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBhV4FcVc/bzRoAABx0u92D/FO7XAMAgAPt0JnurKovS/LcJIer6vvGXU9Ocs6agwEAwCacMZCTfFqSz1oe96Sx/hdJvmmtoQAAYFPOGMjd/fYkb6+q13T3B/ZoJgAA2JgH24N8v0+vqmuSbM3ndPcL1hgKAAA2ZbeB/ItJXpnkVUnuW28cAADYrN0G8ie6+2dWnQQAAPaB3X7M269U1b+sqgur6rz7v1adDAAANmC3e5CPLpc/MNY6yd95ZMcBAIDN2lUgd/fnrT0IAADsB7sK5Kr69p3Wu/u1j+w4AACwWbs9xOLZ4/oTkrwwya1JBDIAAI8quz3E4nvm7ar67CT/dZWJAABgg3b7KRan+r9JLnokBwEAgP1gt8cg/0q2P7UiSc5J8kVJrl1rKAAA2JTdHoP8H8b1TyT5QHfftcI8AACwUbs6xKK7357k/UmelOTcJH+15lAAALApuwrkqvrmJDcluTLJNyd5V1V905qDAQDAJuz2EIsfSfLs7j6ZJFV1OMmvJfmltQYDAIBN2O2nWDzu/jhefPgsngsAAAfGbvcg/2pVvTnJ65fb/yTJf19nJAAA2JwzBnJVfUGSC7r7B6rqG5N8eZJK8s4kr9uD+QAAYE892GESr0jykSTp7jd09/d19/dme+/xK9YdDQAA9t6DBfJWd//OqYvdfXOSrVUmAgCADXqwQH7CGe574iM5CAAA7AcPFsjvrqrvOHWxql6a5JZ1RgIAgM15sE+xeHmS66rqn+avg/hIkk9L8uIV5wIAgI04YyB39z1JnltVX5nki5flE939ltUnAwCADdjV5yB391uTvHXlWQAAYOOcDQ8AAAaBDAAAg0AGAIBBIAMAwLCrX9KDTdk6dmLTI7CytbfxnVdfturrA/DoYw8yAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGA5tegAOtq1jJzY9AgDAI8oeZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGFYL5Kr63Kp6a1XdXlXvq6qXLevnVdUNVXXHcnnuWjMAAMDZWnMP8ieSfH93f1GSS5J8d1U9I8mxJDd290VJblxuAwDAvrBaIHf33d1963L9I0luT/LUJJcnOb487HiSK9aaAQAAztaeHINcVVtJnpXkXUku6O67k+2ITnL+XswAAAC7sXogV9VnJfnlJC/v7r84i+ddVVU3V9XN995773oDAgDAsGogV9Xjsx3Hr+vuNyzL91TVhcv9FyY5udNzu/ua7j7S3UcOHz685pgAAPApa36KRSX52SS3d/dPjLuuT3J0uX40yRvXmgEAAM7WoRVf+3lJvi3Je6vqPcvaDye5Osm1VfXSJB9McuWKMwAAwFlZLZC7+zeS1GnufuFa7wsAAA+HM+kBAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABhWC+SqenVVnayq28baeVV1Q1XdsVyeu9b7AwDAQ7HmHuTXJLn0lLVjSW7s7ouS3LjcBgCAfWO1QO7udyT5k1OWL09yfLl+PMkVa70/AAA8FIf2+P0u6O67k6S7766q80/3wKq6KslVSfL0pz99j8YDODtbx06s+vp3Xn3Zqq8PwAPt21/S6+5ruvtIdx85fPjwpscBAOAxYq8D+Z6qujBJlsuTe/z+AABwRnsdyNcnObpcP5rkjXv8/gAAcEZrfszb65O8M8kXVtVdVfXSJFcneVFV3ZHkRcttAADYN1b7Jb3u/pbT3PXCtd4TAAAern37S3oAALAJAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAw6FNDwCwpq1jJzY9AgAHjD3IAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAACGQ5seYD/bOnZi0yM8LHdefdmmRwAAOHDsQQYAgEEgAwDAIJABAGAQyAAAMAhkAAAYBDIAAAwCGQAABoEMAACDQAYAgEEgAwDAIJABAGAQyAAAMBza9ACsZ+vYiU2PAABw4NiDDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwHBo0wMAcHpbx05seoSH7c6rL1v19df+b7T2/MD+Yw8yAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIDh0KYHAODRbevYiU2PAHBW7EEGAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAAaBDAAAg0AGAIBBIAMAwCCQAQBgOLTpAQBgP9s6dmLTIzwsd1592aZH4IBb+2dgP/4ZtQcZAAAGgQwAAINABgCAQSADAMAgkAEAYBDIAAAwCGQAABgEMgAADAIZAACGjQRyVV1aVb9XVX9QVcc2MQMAAOxkzwO5qs5J8p+TfG2SZyT5lqp6xl7PAQAAO9nEHuTnJPmD7v7D7v6rJL+Q5PINzAEAAA9waAPv+dQk/2fcvivJl576oKq6KslVy82PVtXv7cFsj2VPSfLHmx6Cs2a7HVy23cF04LZb/fimJ9gXDtx2eyx5kD+ja2+7v73T4iYCuXZY6wcsdF+T5Jr1xyFJqurm7j6y6Tk4O7bbwWXbHUy228Fkux1cm9p2mzjE4q4knztuPy3JhzYwBwAAPMAmAvndSS6qqs+rqk9L8pIk129gDgAAeIA9P8Siuz9RVf8qyZuTnJPk1d39vr2egwdwOMvBZLsdXLbdwWS7HUy228G1kW1X3Q84/BcAAB6znEkPAAAGgQwAAINAfoypqiur6n1V9cmqOnLKfT+0nP7796rqa8b6P6iq9y73/WRV7fRRfewxp2zfv6rq1VV1sqpuG2vnVdUNVXXHcnnuuG/Hnz32VlV9blW9tapuX/6efNmybtvtc1X1hKq6qap+e9l2P7as23YHQFWdU1W/VVVvWm5vfLsJ5Mee25J8Y5J3zMXldN8vSfL3klya5KeX04Inyc9k+6QtFy1fl+7ZtOzIKdv3vdfkgT8nx5Lc2N0XJblxuf1gP3vsrU8k+f7u/qIklyT57mX72Hb738eTvKC7n5nk4iSXVtUlse0OipcluX3c3vh2E8iPMd19e3fvdFbCy5P8Qnd/vLv/KMkfJHlOVV2Y5Mnd/c7e/o3O1ya5Yu8m5jScsn0f6+53JPmTU5YvT3J8uX48f/1ztOPP3l7Myd/U3Xd3963L9Y9k+3/YT41tt+/1to8uNx+/fHVsu32vqp6W5LIkrxrLG99uApn77XQK8KcuX3ftsM5mnW57sX9d0N13J9shluT8Zd223IeqaivJs5K8K7bdgbD8M/17kpxMckN323YHwyuS/GCST461jW+3TZxqmpVV1a8l+Vs73PUj3f3G0z1th7U+wzqbZbs8etiW+0xVfVaSX07y8u7+izP82oVtt490931JLq6qz0lyXVV98RkebtvtA1X19UlOdvctVfX83Txlh7VVtptAfhTq7q96CE873SnA71qun7rOZjll+8FzT1Vd2N13L4cunVzWbct9pKoen+04fl13v2FZtu0OkO7+s6p6W7aPUbXt9rfnJfmGqvq6JE9I8uSq+rnsg+3mEAvud32Sl1TVp1fV52X7l/FuWv5p4yNVdcny6RXfnuR0e6HZO07ZfvBcn+Tocv1o/vrnaMefvQ3M95i3/B33s0lu7+6fGHfZdvtcVR1e9hynqp6Y5KuSvD+23b7W3T/U3U/r7q1s/3/sLd39rdkH280e5MeYqnpxkp9KcjjJiap6T3d/TXe/r6quTfK72f5N7u9e/rkqSb4r27+V/8Qk/2P5YoOcsn1/q6rXJ3l+kqdU1V1JfjTJ1UmuraqXJvlgkiuT5EF+9thbz0vybUneuxzLmiQ/HNvuILgwyfHlEw0el+Ta7n5TVb0ztt1BtPGfOaeaBgCAwSEWAAAwCGQAABgEMgAADAIZAAAGgQwAAINABniIquq+qnpPVd1WVb9YVZ/xMF7rNVX1Tcv1V1XVM87w2OdX1XPH7e+sqm9/qO89Xmerqj62fE/3fz3s1wU4aHwOMsBD97HuvjhJqup1Sb4zyadOMFFV5zyUz+js7n/+IA95fpKPJvlfy+NfebbvcQb/+/7v6XRO/b52830uJ+Go7v7kIzMmwHrsQQZ4ZPx6ki9Y9u6+tap+PtsnnDinqv59Vb27qn6nqv5Fsh2MVfWfqup3q+pEkvPvf6GqeltVHVmuX1pVt1bVb1fVjVW1le0Q/95lD+9XVNW/rap/vTz+4qr6zeW9rquqc8dr/nhV3VRVv19VX3E231xVfbSq/l1VvSvJl+1w+/uWPem3VdXLl+dsVdXtVfXTSW7N3zxFLMC+JZABHqaqOpTka5O8d1l6TpIf6e5nJHlpkj/v7mcneXaS71hOkfriJF+Y5O8n+Y4kz93hdQ8n+S9J/nF3PzPJld19Z5JXJvmP3X1xd//6KU97bZJ/091fsszzo+O+Q939nCQvP2V9+vxTDrG4P6Q/M8lt3f2l3f0b83aSjyX5Z0m+NMkly/f4rOV5X5jktd39rO7+wGn/IwLsIw6xAHjonjhOSfzrSX4226F7U3f/0bL+1Um+5P7ji5N8dpKLkvzDJK9fDk34UFW9ZYfXvyTJO+5/re7+kzMNU1WfneRzuvvty9LxJL84HvKG5fKWJFuneZnTHWJxX5JfPs3tL09yXXf/5TLHG5J8RZLrk3ygu3/zTHMD7DcCGeCh+9ipMbl9qG3+ci4l+Z7ufvMpj/u6JP0gr1+7eMzZ+PhyeV/O/u///3fKccbzdp3heX95hvsA9iWHWACs681JvquqHp8kVfV3q+ozk7wjyUuWY5QvTPKVOzz3nUn+0XJIRqrqvGX9I0medOqDu/vPk/zpOCzi25K8/dTHreAdSa6oqs9YvrcXZ3uPOsCBZA8ywLpele3DGW5dPsnh3iRXJLkuyQuyfZzw72eHkO3ue6vqqiRvqKrHJTmZ5EVJfiXJL1XV5Um+55SnHU3yyuUj5/4w28cGn43PH4eNJMmru/snz/SE7r61ql6T5KZl6VXd/VvLLxQCHDjV/Uj+6x0AABxsDrEAAIBBIAMAwCCQAQBgEMgAADAIZAAAGAQyAAAMAhkAAIb/D0hAYsYxi5IBAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib\n",
    "\n",
    "error = y_pred - y_test\n",
    "plt.hist(error, bins=25)\n",
    "plt.xlabel('Prediction Error')\n",
    "_ = plt.ylabel('Count')\n",
    "\n",
    "fig = matplotlib.pyplot.gcf()\n",
    "fig.set_size_inches(10,10)\n",
    "fig.savefig('pred_error_gbr.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 277,
   "id": "12c560d4-e963-4498-b68a-49d42682a2fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate_predictions(predictions, y_test, outliers):\n",
    "    ratio = []\n",
    "    differences = []\n",
    "    for pred in range(len(y_test)):\n",
    "        ratio.append((y_test[pred]/predictions[pred])-1)\n",
    "        differences.append(abs(y_test[pred]- predictions[pred]))\n",
    "    \n",
    "    \n",
    "    n_outliers = int(len(differences) * outliers)\n",
    "    outliers = pd.Series(differences).astype(float).nlargest(n_outliers)\n",
    "    \n",
    "    return ratio, differences, outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "id": "64962f2c-b30f-4d10-b21a-26abc181f437",
   "metadata": {},
   "outputs": [],
   "source": [
    "ratio, differences, outliers = evaluate_predictions(y_pred, y_test, 0.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "id": "a93e1575-cfe4-4e75-9f93-f898681b388e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "99     655.579998\n",
       "234    526.530398\n",
       "dtype: float64"
      ]
     },
     "execution_count": 279,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for index in outliers.index: \n",
    "    outliers[index] = y_pred[index]\n",
    "outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "id": "37e1f5e1-2131-427c-8c95-a367e279fb2e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots()\n",
    "fig.set_figheight(4)\n",
    "fig.set_figwidth(15)\n",
    "\n",
    "axs.plot(y_pred,color='red', label='Predicted')\n",
    "axs.plot(y_test,color='blue', label='Actual')\n",
    "#axs.scatter(outliers.index,outliers, color='green', linewidth=5.0, label='Anomalies')\n",
    "plt.xlabel('Timestamp')\n",
    "plt.ylabel('Scaled number of passengers')\n",
    "plt.legend(loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "951f7f3f-a21e-4f7a-aa17-8057022d6f5d",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Simple ANN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "id": "f3c495a9-10cc-4545-ae02-58ac010e9d01",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.layers import LSTM\n",
    "from keras.models import Sequential\n",
    "#from keras.layers.wrappers import TimeDistributed\n",
    "from keras.layers.core import Dense, Activation, Dropout\n",
    "from tensorflow.keras import models, layers\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "import tensorflow_addons as tfa\n",
    "import tensorflow as tf\n",
    "import time\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "id": "0e7ec30f-fe23-4319-b3ae-abf9e0b85182",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_relu1 = models.Sequential([\n",
    "    layers.Dense(256, activation='relu'),\n",
    "    layers.Dense(256, activation='relu'),\n",
    "    layers.Dense(128, activation='relu'),\n",
    "    layers.Dense(128, activation='relu'),\n",
    "    layers.Dense(64, activation='relu'),\n",
    "    layers.Dense(64, activation='relu'),\n",
    "    layers.Dense(1, activation='relu')])\n",
    "\n",
    "model_relu1.compile(metrics=[tfa.metrics.RSquare()],\n",
    "                 loss='mean_squared_error',\n",
    "                 optimizer=tf.keras.optimizers.Adam(0.001))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "id": "c1b6f051-3939-4a69-bda9-9ebd8a63b9ff",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/400\n",
      "8/8 [==============================] - 2s 49ms/step - loss: 391658.1562 - r_square: -37.0906 - val_loss: 419726.7500 - val_r_square: -47.8819\n",
      "Epoch 2/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 377826.3125 - r_square: -35.7458 - val_loss: 377310.1250 - val_r_square: -42.9420\n",
      "Epoch 3/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 284478.7500 - r_square: -26.6671 - val_loss: 155735.9688 - val_r_square: -17.1372\n",
      "Epoch 4/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 74956.7891 - r_square: -6.2900 - val_loss: 34792.3867 - val_r_square: -3.0520\n",
      "Epoch 5/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 28662.1855 - r_square: -1.7876 - val_loss: 38418.9688 - val_r_square: -3.4743\n",
      "Epoch 6/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 28553.0996 - r_square: -1.7769 - val_loss: 19760.3809 - val_r_square: -1.3013\n",
      "Epoch 7/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 17569.7422 - r_square: -0.7088 - val_loss: 13138.9082 - val_r_square: -0.5302\n",
      "Epoch 8/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 13933.9561 - r_square: -0.3551 - val_loss: 13658.7900 - val_r_square: -0.5907\n",
      "Epoch 9/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 12295.0801 - r_square: -0.1958 - val_loss: 10812.0762 - val_r_square: -0.2592\n",
      "Epoch 10/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 10724.7461 - r_square: -0.0430 - val_loss: 11578.3213 - val_r_square: -0.3484\n",
      "Epoch 11/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 10216.2559 - r_square: 0.0064 - val_loss: 10228.9502 - val_r_square: -0.1913\n",
      "Epoch 12/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 9374.9434 - r_square: 0.0882 - val_loss: 9750.9541 - val_r_square: -0.1356\n",
      "Epoch 13/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 8979.7256 - r_square: 0.1267 - val_loss: 9642.5430 - val_r_square: -0.1230\n",
      "Epoch 14/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 8666.6338 - r_square: 0.1571 - val_loss: 9537.9141 - val_r_square: -0.1108\n",
      "Epoch 15/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 8340.5869 - r_square: 0.1888 - val_loss: 8797.6572 - val_r_square: -0.0246\n",
      "Epoch 16/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 8072.0366 - r_square: 0.2150 - val_loss: 8889.0742 - val_r_square: -0.0352\n",
      "Epoch 17/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 7799.6914 - r_square: 0.2414 - val_loss: 8464.7109 - val_r_square: 0.0142\n",
      "Epoch 18/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 7600.9941 - r_square: 0.2608 - val_loss: 8641.6299 - val_r_square: -0.0064\n",
      "Epoch 19/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 7436.6431 - r_square: 0.2767 - val_loss: 8268.4072 - val_r_square: 0.0371\n",
      "Epoch 20/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 7305.8931 - r_square: 0.2895 - val_loss: 8386.9941 - val_r_square: 0.0232\n",
      "Epoch 21/400\n",
      "8/8 [==============================] - 0s 15ms/step - loss: 7108.2461 - r_square: 0.3087 - val_loss: 7905.9756 - val_r_square: 0.0793\n",
      "Epoch 22/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 6999.0469 - r_square: 0.3193 - val_loss: 8190.6279 - val_r_square: 0.0461\n",
      "Epoch 23/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 6973.7417 - r_square: 0.3218 - val_loss: 7727.4683 - val_r_square: 0.1000\n",
      "Epoch 24/400\n",
      "8/8 [==============================] - 0s 14ms/step - loss: 6798.0933 - r_square: 0.3389 - val_loss: 8135.8804 - val_r_square: 0.0525\n",
      "Epoch 25/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 6773.4111 - r_square: 0.3413 - val_loss: 7188.7783 - val_r_square: 0.1628\n",
      "Epoch 26/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 6744.2231 - r_square: 0.3441 - val_loss: 8115.5942 - val_r_square: 0.0548\n",
      "Epoch 27/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 6692.3115 - r_square: 0.3491 - val_loss: 7545.1626 - val_r_square: 0.1213\n",
      "Epoch 28/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 6818.4038 - r_square: 0.3369 - val_loss: 6878.8691 - val_r_square: 0.1989\n",
      "Epoch 29/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 6420.8301 - r_square: 0.3755 - val_loss: 7565.3691 - val_r_square: 0.1189\n",
      "Epoch 30/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 6312.3979 - r_square: 0.3861 - val_loss: 6821.3809 - val_r_square: 0.2056\n",
      "Epoch 31/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 6394.6909 - r_square: 0.3781 - val_loss: 6861.2295 - val_r_square: 0.2009\n",
      "Epoch 32/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 6474.5708 - r_square: 0.3703 - val_loss: 7427.9937 - val_r_square: 0.1349\n",
      "Epoch 33/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 6199.4756 - r_square: 0.3971 - val_loss: 6821.2554 - val_r_square: 0.2056\n",
      "Epoch 34/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 6117.4878 - r_square: 0.4050 - val_loss: 6663.9312 - val_r_square: 0.2239\n",
      "Epoch 35/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 6111.3086 - r_square: 0.4056 - val_loss: 7089.8931 - val_r_square: 0.1743\n",
      "Epoch 36/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 6016.6748 - r_square: 0.4148 - val_loss: 6828.8242 - val_r_square: 0.2047\n",
      "Epoch 37/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 6052.9795 - r_square: 0.4113 - val_loss: 6743.3745 - val_r_square: 0.2147\n",
      "Epoch 38/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 5942.0601 - r_square: 0.4221 - val_loss: 6493.9448 - val_r_square: 0.2437\n",
      "Epoch 39/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5934.9028 - r_square: 0.4228 - val_loss: 7181.2456 - val_r_square: 0.1637\n",
      "Epoch 40/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 5915.4180 - r_square: 0.4247 - val_loss: 6539.7280 - val_r_square: 0.2384\n",
      "Epoch 41/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 5871.3511 - r_square: 0.4290 - val_loss: 6460.2188 - val_r_square: 0.2476\n",
      "Epoch 42/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5838.8276 - r_square: 0.4321 - val_loss: 6804.7598 - val_r_square: 0.2075\n",
      "Epoch 43/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5742.9854 - r_square: 0.4415 - val_loss: 6336.9116 - val_r_square: 0.2620\n",
      "Epoch 44/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 5787.3945 - r_square: 0.4371 - val_loss: 6567.8115 - val_r_square: 0.2351\n",
      "Epoch 45/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5754.1240 - r_square: 0.4404 - val_loss: 6695.3940 - val_r_square: 0.2202\n",
      "Epoch 46/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5888.2627 - r_square: 0.4273 - val_loss: 6294.0771 - val_r_square: 0.2670\n",
      "Epoch 47/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5770.3330 - r_square: 0.4388 - val_loss: 6002.4136 - val_r_square: 0.3010\n",
      "Epoch 48/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5761.2920 - r_square: 0.4397 - val_loss: 6331.7378 - val_r_square: 0.2626\n",
      "Epoch 49/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5593.8223 - r_square: 0.4560 - val_loss: 6183.6250 - val_r_square: 0.2798\n",
      "Epoch 50/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5701.1987 - r_square: 0.4455 - val_loss: 6627.7837 - val_r_square: 0.2281\n",
      "Epoch 51/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5649.7280 - r_square: 0.4505 - val_loss: 6112.0679 - val_r_square: 0.2882\n",
      "Epoch 52/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5502.7051 - r_square: 0.4648 - val_loss: 6153.4028 - val_r_square: 0.2834\n",
      "Epoch 53/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5482.6694 - r_square: 0.4668 - val_loss: 7051.1431 - val_r_square: 0.1788\n",
      "Epoch 54/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 5670.4438 - r_square: 0.4485 - val_loss: 6450.9180 - val_r_square: 0.2487\n",
      "Epoch 55/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5801.2583 - r_square: 0.4358 - val_loss: 5910.8208 - val_r_square: 0.3116\n",
      "Epoch 56/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 5606.3604 - r_square: 0.4548 - val_loss: 5728.7705 - val_r_square: 0.3328\n",
      "Epoch 57/400\n",
      "8/8 [==============================] - 1s 82ms/step - loss: 5644.6079 - r_square: 0.4510 - val_loss: 6201.8745 - val_r_square: 0.2777\n",
      "Epoch 58/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5428.3130 - r_square: 0.4721 - val_loss: 6174.4194 - val_r_square: 0.2809\n",
      "Epoch 59/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 5313.0435 - r_square: 0.4833 - val_loss: 5985.5195 - val_r_square: 0.3029\n",
      "Epoch 60/400\n",
      "8/8 [==============================] - 0s 28ms/step - loss: 5331.4248 - r_square: 0.4815 - val_loss: 5673.0054 - val_r_square: 0.3393\n",
      "Epoch 61/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5252.3672 - r_square: 0.4892 - val_loss: 6753.2808 - val_r_square: 0.2135\n",
      "Epoch 62/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 5509.5083 - r_square: 0.4642 - val_loss: 6277.5771 - val_r_square: 0.2689\n",
      "Epoch 63/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5347.7900 - r_square: 0.4799 - val_loss: 5543.9731 - val_r_square: 0.3543\n",
      "Epoch 64/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5216.5366 - r_square: 0.4927 - val_loss: 5651.7734 - val_r_square: 0.3418\n",
      "Epoch 65/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 5531.9541 - r_square: 0.4620 - val_loss: 7207.3267 - val_r_square: 0.1606\n",
      "Epoch 66/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5291.4893 - r_square: 0.4854 - val_loss: 5832.1777 - val_r_square: 0.3208\n",
      "Epoch 67/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 5186.2393 - r_square: 0.4956 - val_loss: 5585.8071 - val_r_square: 0.3495\n",
      "Epoch 68/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 5113.9272 - r_square: 0.5026 - val_loss: 5589.8579 - val_r_square: 0.3490\n",
      "Epoch 69/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 5047.6406 - r_square: 0.5091 - val_loss: 6139.2241 - val_r_square: 0.2850\n",
      "Epoch 70/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 5072.5405 - r_square: 0.5067 - val_loss: 5909.2817 - val_r_square: 0.3118\n",
      "Epoch 71/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 5027.5171 - r_square: 0.5110 - val_loss: 5431.7437 - val_r_square: 0.3674\n",
      "Epoch 72/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 5151.8179 - r_square: 0.4990 - val_loss: 5818.3149 - val_r_square: 0.3224\n",
      "Epoch 73/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4900.2510 - r_square: 0.5234 - val_loss: 5238.0029 - val_r_square: 0.3900\n",
      "Epoch 74/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 4949.9482 - r_square: 0.5186 - val_loss: 5577.7314 - val_r_square: 0.3504\n",
      "Epoch 75/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4913.0151 - r_square: 0.5222 - val_loss: 5501.5176 - val_r_square: 0.3593\n",
      "Epoch 76/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 4908.8809 - r_square: 0.5226 - val_loss: 5727.1279 - val_r_square: 0.3330\n",
      "Epoch 77/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4877.9565 - r_square: 0.5256 - val_loss: 5658.9312 - val_r_square: 0.3410\n",
      "Epoch 78/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4858.8750 - r_square: 0.5275 - val_loss: 5359.4502 - val_r_square: 0.3758\n",
      "Epoch 79/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4850.2173 - r_square: 0.5283 - val_loss: 5610.8755 - val_r_square: 0.3466\n",
      "Epoch 80/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 4925.9517 - r_square: 0.5209 - val_loss: 5252.7666 - val_r_square: 0.3883\n",
      "Epoch 81/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 5013.6685 - r_square: 0.5124 - val_loss: 5404.3936 - val_r_square: 0.3706\n",
      "Epoch 82/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 5054.7666 - r_square: 0.5084 - val_loss: 4969.3027 - val_r_square: 0.4213\n",
      "Epoch 83/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 4996.2651 - r_square: 0.5141 - val_loss: 5299.6704 - val_r_square: 0.3828\n",
      "Epoch 84/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4712.1021 - r_square: 0.5417 - val_loss: 5276.9229 - val_r_square: 0.3854\n",
      "Epoch 85/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 4655.0078 - r_square: 0.5473 - val_loss: 5362.1362 - val_r_square: 0.3755\n",
      "Epoch 86/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 4665.7212 - r_square: 0.5462 - val_loss: 4895.4556 - val_r_square: 0.4299\n",
      "Epoch 87/400\n",
      "8/8 [==============================] - 0s 15ms/step - loss: 5124.3472 - r_square: 0.5016 - val_loss: 5344.9585 - val_r_square: 0.3775\n",
      "Epoch 88/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4741.4707 - r_square: 0.5389 - val_loss: 5162.9810 - val_r_square: 0.3987\n",
      "Epoch 89/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4692.9409 - r_square: 0.5436 - val_loss: 5001.9185 - val_r_square: 0.4175\n",
      "Epoch 90/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4661.1699 - r_square: 0.5467 - val_loss: 5304.4092 - val_r_square: 0.3822\n",
      "Epoch 91/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4527.0820 - r_square: 0.5597 - val_loss: 5153.0239 - val_r_square: 0.3999\n",
      "Epoch 92/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4582.7769 - r_square: 0.5543 - val_loss: 5098.4795 - val_r_square: 0.4062\n",
      "Epoch 93/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4552.2144 - r_square: 0.5573 - val_loss: 5397.3281 - val_r_square: 0.3714\n",
      "Epoch 94/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4616.5640 - r_square: 0.5510 - val_loss: 5802.3008 - val_r_square: 0.3243\n",
      "Epoch 95/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 4538.6904 - r_square: 0.5586 - val_loss: 5973.5000 - val_r_square: 0.3043\n",
      "Epoch 96/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4517.3818 - r_square: 0.5607 - val_loss: 5129.9814 - val_r_square: 0.4026\n",
      "Epoch 97/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4485.1523 - r_square: 0.5638 - val_loss: 4979.2251 - val_r_square: 0.4201\n",
      "Epoch 98/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 4412.5669 - r_square: 0.5709 - val_loss: 4783.7666 - val_r_square: 0.4429\n",
      "Epoch 99/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4315.4395 - r_square: 0.5803 - val_loss: 4583.6616 - val_r_square: 0.4662\n",
      "Epoch 100/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 4481.1133 - r_square: 0.5642 - val_loss: 5420.1606 - val_r_square: 0.3688\n",
      "Epoch 101/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4265.8164 - r_square: 0.5851 - val_loss: 5417.9194 - val_r_square: 0.3690\n",
      "Epoch 102/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4380.1812 - r_square: 0.5740 - val_loss: 4966.1748 - val_r_square: 0.4216\n",
      "Epoch 103/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4335.2378 - r_square: 0.5784 - val_loss: 4503.7695 - val_r_square: 0.4755\n",
      "Epoch 104/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 4516.2183 - r_square: 0.5608 - val_loss: 4512.6279 - val_r_square: 0.4745\n",
      "Epoch 105/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4250.5576 - r_square: 0.5866 - val_loss: 4638.9248 - val_r_square: 0.4597\n",
      "Epoch 106/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 4265.6753 - r_square: 0.5851 - val_loss: 4575.3076 - val_r_square: 0.4672\n",
      "Epoch 107/400\n",
      "8/8 [==============================] - 0s 14ms/step - loss: 4351.7974 - r_square: 0.5768 - val_loss: 4511.9248 - val_r_square: 0.4745\n",
      "Epoch 108/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4320.4624 - r_square: 0.5798 - val_loss: 4788.4990 - val_r_square: 0.4423\n",
      "Epoch 109/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4206.3789 - r_square: 0.5909 - val_loss: 5613.3940 - val_r_square: 0.3463\n",
      "Epoch 110/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4564.9766 - r_square: 0.5560 - val_loss: 4912.1987 - val_r_square: 0.4279\n",
      "Epoch 111/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4244.1548 - r_square: 0.5872 - val_loss: 4566.7808 - val_r_square: 0.4681\n",
      "Epoch 112/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4055.1904 - r_square: 0.6056 - val_loss: 4451.0146 - val_r_square: 0.4816\n",
      "Epoch 113/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4061.7058 - r_square: 0.6050 - val_loss: 4294.6929 - val_r_square: 0.4998\n",
      "Epoch 114/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4095.5549 - r_square: 0.6017 - val_loss: 4605.9116 - val_r_square: 0.4636\n",
      "Epoch 115/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3946.5212 - r_square: 0.6162 - val_loss: 5092.7666 - val_r_square: 0.4069\n",
      "Epoch 116/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 4131.6348 - r_square: 0.5982 - val_loss: 5348.1304 - val_r_square: 0.3772\n",
      "Epoch 117/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 4415.2070 - r_square: 0.5706 - val_loss: 4764.6660 - val_r_square: 0.4451\n",
      "Epoch 118/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4041.7983 - r_square: 0.6069 - val_loss: 4975.4277 - val_r_square: 0.4206\n",
      "Epoch 119/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3823.8843 - r_square: 0.6281 - val_loss: 4789.6362 - val_r_square: 0.4422\n",
      "Epoch 120/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3968.3606 - r_square: 0.6141 - val_loss: 4512.5176 - val_r_square: 0.4745\n",
      "Epoch 121/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3894.3013 - r_square: 0.6213 - val_loss: 4221.3623 - val_r_square: 0.5084\n",
      "Epoch 122/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 4004.8611 - r_square: 0.6105 - val_loss: 4506.0908 - val_r_square: 0.4752\n",
      "Epoch 123/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3958.2314 - r_square: 0.6150 - val_loss: 4185.0649 - val_r_square: 0.5126\n",
      "Epoch 124/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 4070.7361 - r_square: 0.6041 - val_loss: 4244.3892 - val_r_square: 0.5057\n",
      "Epoch 125/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 4154.5713 - r_square: 0.5959 - val_loss: 4551.3120 - val_r_square: 0.4699\n",
      "Epoch 126/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3800.0396 - r_square: 0.6304 - val_loss: 4649.3296 - val_r_square: 0.4585\n",
      "Epoch 127/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 3770.6050 - r_square: 0.6333 - val_loss: 4150.2461 - val_r_square: 0.5167\n",
      "Epoch 128/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3812.4006 - r_square: 0.6292 - val_loss: 4377.6636 - val_r_square: 0.4902\n",
      "Epoch 129/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3962.8689 - r_square: 0.6146 - val_loss: 4367.2183 - val_r_square: 0.4914\n",
      "Epoch 130/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3830.5483 - r_square: 0.6275 - val_loss: 3983.9546 - val_r_square: 0.5360\n",
      "Epoch 131/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 4212.5493 - r_square: 0.5903 - val_loss: 4944.1680 - val_r_square: 0.4242\n",
      "Epoch 132/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 4011.5378 - r_square: 0.6099 - val_loss: 4380.4434 - val_r_square: 0.4898\n",
      "Epoch 133/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3760.5742 - r_square: 0.6343 - val_loss: 4910.6812 - val_r_square: 0.4281\n",
      "Epoch 134/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3818.6736 - r_square: 0.6286 - val_loss: 5634.8423 - val_r_square: 0.3438\n",
      "Epoch 135/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3950.0386 - r_square: 0.6158 - val_loss: 5351.3594 - val_r_square: 0.3768\n",
      "Epoch 136/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3798.3882 - r_square: 0.6306 - val_loss: 5161.5630 - val_r_square: 0.3989\n",
      "Epoch 137/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3744.5881 - r_square: 0.6358 - val_loss: 4414.1426 - val_r_square: 0.4859\n",
      "Epoch 138/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3680.2451 - r_square: 0.6421 - val_loss: 4345.2646 - val_r_square: 0.4939\n",
      "Epoch 139/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3651.1602 - r_square: 0.6449 - val_loss: 4841.1333 - val_r_square: 0.4362\n",
      "Epoch 140/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3789.8359 - r_square: 0.6314 - val_loss: 4641.0830 - val_r_square: 0.4595\n",
      "Epoch 141/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3758.6160 - r_square: 0.6345 - val_loss: 4214.3096 - val_r_square: 0.5092\n",
      "Epoch 142/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3565.6982 - r_square: 0.6532 - val_loss: 4331.1060 - val_r_square: 0.4956\n",
      "Epoch 143/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 3653.8853 - r_square: 0.6446 - val_loss: 3966.8411 - val_r_square: 0.5380\n",
      "Epoch 144/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3647.8037 - r_square: 0.6452 - val_loss: 4318.6064 - val_r_square: 0.4970\n",
      "Epoch 145/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3723.4722 - r_square: 0.6379 - val_loss: 4746.4810 - val_r_square: 0.4472\n",
      "Epoch 146/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3646.2620 - r_square: 0.6454 - val_loss: 4091.1567 - val_r_square: 0.5235\n",
      "Epoch 147/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3704.5391 - r_square: 0.6397 - val_loss: 4261.2827 - val_r_square: 0.5037\n",
      "Epoch 148/400\n",
      "8/8 [==============================] - 0s 14ms/step - loss: 3584.2908 - r_square: 0.6514 - val_loss: 4022.6758 - val_r_square: 0.5315\n",
      "Epoch 149/400\n",
      "8/8 [==============================] - 0s 14ms/step - loss: 3670.1636 - r_square: 0.6431 - val_loss: 3989.8875 - val_r_square: 0.5353\n",
      "Epoch 150/400\n",
      "8/8 [==============================] - 0s 15ms/step - loss: 3536.8684 - r_square: 0.6560 - val_loss: 4044.2703 - val_r_square: 0.5290\n",
      "Epoch 151/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3528.9338 - r_square: 0.6568 - val_loss: 4243.1548 - val_r_square: 0.5058\n",
      "Epoch 152/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3676.7205 - r_square: 0.6424 - val_loss: 4193.2080 - val_r_square: 0.5117\n",
      "Epoch 153/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3621.3687 - r_square: 0.6478 - val_loss: 4285.0166 - val_r_square: 0.5010\n",
      "Epoch 154/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3622.8374 - r_square: 0.6477 - val_loss: 4967.4082 - val_r_square: 0.4215\n",
      "Epoch 155/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3582.1799 - r_square: 0.6516 - val_loss: 6558.5723 - val_r_square: 0.2362\n",
      "Epoch 156/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3934.2659 - r_square: 0.6174 - val_loss: 4882.4185 - val_r_square: 0.4314\n",
      "Epoch 157/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3434.3826 - r_square: 0.6660 - val_loss: 4646.6636 - val_r_square: 0.4588\n",
      "Epoch 158/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3543.5549 - r_square: 0.6554 - val_loss: 4215.9131 - val_r_square: 0.5090\n",
      "Epoch 159/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3552.3689 - r_square: 0.6545 - val_loss: 4575.1523 - val_r_square: 0.4672\n",
      "Epoch 160/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3475.0142 - r_square: 0.6620 - val_loss: 4000.2205 - val_r_square: 0.5341\n",
      "Epoch 161/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3937.8569 - r_square: 0.6170 - val_loss: 3929.9897 - val_r_square: 0.5423\n",
      "Epoch 162/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3419.7090 - r_square: 0.6674 - val_loss: 4626.3379 - val_r_square: 0.4612\n",
      "Epoch 163/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3360.3289 - r_square: 0.6732 - val_loss: 4729.5723 - val_r_square: 0.4492\n",
      "Epoch 164/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3422.4648 - r_square: 0.6671 - val_loss: 4601.0186 - val_r_square: 0.4642\n",
      "Epoch 165/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 3416.9075 - r_square: 0.6677 - val_loss: 5745.4448 - val_r_square: 0.3309\n",
      "Epoch 166/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3501.8936 - r_square: 0.6594 - val_loss: 4341.5132 - val_r_square: 0.4944\n",
      "Epoch 167/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3334.5825 - r_square: 0.6757 - val_loss: 4643.8398 - val_r_square: 0.4592\n",
      "Epoch 168/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3255.7983 - r_square: 0.6834 - val_loss: 4957.7236 - val_r_square: 0.4226\n",
      "Epoch 169/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3266.0415 - r_square: 0.6824 - val_loss: 4412.8477 - val_r_square: 0.4861\n",
      "Epoch 170/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3133.0696 - r_square: 0.6953 - val_loss: 4080.9241 - val_r_square: 0.5247\n",
      "Epoch 171/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3180.6497 - r_square: 0.6907 - val_loss: 4144.4819 - val_r_square: 0.5173\n",
      "Epoch 172/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3186.3660 - r_square: 0.6901 - val_loss: 3764.1074 - val_r_square: 0.5616\n",
      "Epoch 173/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 3149.0942 - r_square: 0.6937 - val_loss: 3847.8816 - val_r_square: 0.5519\n",
      "Epoch 174/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3331.2947 - r_square: 0.6760 - val_loss: 4745.0327 - val_r_square: 0.4474\n",
      "Epoch 175/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3397.4673 - r_square: 0.6696 - val_loss: 4272.8359 - val_r_square: 0.5024\n",
      "Epoch 176/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3217.4399 - r_square: 0.6871 - val_loss: 4508.9019 - val_r_square: 0.4749\n",
      "Epoch 177/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3116.8782 - r_square: 0.6969 - val_loss: 4217.9106 - val_r_square: 0.5088\n",
      "Epoch 178/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 3301.5464 - r_square: 0.6789 - val_loss: 4264.8486 - val_r_square: 0.5033\n",
      "Epoch 179/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3555.3672 - r_square: 0.6542 - val_loss: 3961.1311 - val_r_square: 0.5387\n",
      "Epoch 180/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3787.9089 - r_square: 0.6316 - val_loss: 3974.4121 - val_r_square: 0.5371\n",
      "Epoch 181/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3595.2065 - r_square: 0.6503 - val_loss: 4272.9927 - val_r_square: 0.5024\n",
      "Epoch 182/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3520.3538 - r_square: 0.6576 - val_loss: 3823.1577 - val_r_square: 0.5548\n",
      "Epoch 183/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3868.6211 - r_square: 0.6238 - val_loss: 4519.8794 - val_r_square: 0.4736\n",
      "Epoch 184/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3542.7346 - r_square: 0.6554 - val_loss: 4041.3538 - val_r_square: 0.5293\n",
      "Epoch 185/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3565.4265 - r_square: 0.6532 - val_loss: 3847.4167 - val_r_square: 0.5519\n",
      "Epoch 186/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 3259.3884 - r_square: 0.6830 - val_loss: 4675.9194 - val_r_square: 0.4554\n",
      "Epoch 187/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3258.4048 - r_square: 0.6831 - val_loss: 4495.7412 - val_r_square: 0.4764\n",
      "Epoch 188/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3227.3472 - r_square: 0.6861 - val_loss: 4874.4180 - val_r_square: 0.4323\n",
      "Epoch 189/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3200.6343 - r_square: 0.6887 - val_loss: 4900.4878 - val_r_square: 0.4293\n",
      "Epoch 190/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3249.8386 - r_square: 0.6839 - val_loss: 4739.7334 - val_r_square: 0.4480\n",
      "Epoch 191/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3033.1580 - r_square: 0.7050 - val_loss: 4253.4961 - val_r_square: 0.5046\n",
      "Epoch 192/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3149.0288 - r_square: 0.6937 - val_loss: 4644.4150 - val_r_square: 0.4591\n",
      "Epoch 193/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3180.6187 - r_square: 0.6907 - val_loss: 3678.7996 - val_r_square: 0.5716\n",
      "Epoch 194/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3159.7522 - r_square: 0.6927 - val_loss: 4178.5464 - val_r_square: 0.5134\n",
      "Epoch 195/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3048.2695 - r_square: 0.7035 - val_loss: 4101.0933 - val_r_square: 0.5224\n",
      "Epoch 196/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3056.0085 - r_square: 0.7028 - val_loss: 4444.4316 - val_r_square: 0.4824\n",
      "Epoch 197/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3359.5046 - r_square: 0.6733 - val_loss: 4105.1753 - val_r_square: 0.5219\n",
      "Epoch 198/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3227.0171 - r_square: 0.6862 - val_loss: 4164.2578 - val_r_square: 0.5150\n",
      "Epoch 199/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3335.1882 - r_square: 0.6756 - val_loss: 4701.8843 - val_r_square: 0.4524\n",
      "Epoch 200/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3299.0696 - r_square: 0.6791 - val_loss: 3879.6611 - val_r_square: 0.5482\n",
      "Epoch 201/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2898.3611 - r_square: 0.7181 - val_loss: 4301.5483 - val_r_square: 0.4990\n",
      "Epoch 202/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2920.0356 - r_square: 0.7160 - val_loss: 3726.7893 - val_r_square: 0.5660\n",
      "Epoch 203/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 3167.3093 - r_square: 0.6920 - val_loss: 4200.5054 - val_r_square: 0.5108\n",
      "Epoch 204/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3037.3835 - r_square: 0.7046 - val_loss: 3837.9360 - val_r_square: 0.5530\n",
      "Epoch 205/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3026.7666 - r_square: 0.7056 - val_loss: 3881.0693 - val_r_square: 0.5480\n",
      "Epoch 206/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2964.5015 - r_square: 0.7117 - val_loss: 4509.6948 - val_r_square: 0.4748\n",
      "Epoch 207/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3795.2981 - r_square: 0.6309 - val_loss: 4938.3003 - val_r_square: 0.4249\n",
      "Epoch 208/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3543.6384 - r_square: 0.6554 - val_loss: 3805.9785 - val_r_square: 0.5568\n",
      "Epoch 209/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 2929.5964 - r_square: 0.7151 - val_loss: 4024.0583 - val_r_square: 0.5314\n",
      "Epoch 210/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3082.5601 - r_square: 0.7002 - val_loss: 3905.3237 - val_r_square: 0.5452\n",
      "Epoch 211/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2891.3435 - r_square: 0.7188 - val_loss: 3877.0708 - val_r_square: 0.5485\n",
      "Epoch 212/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2810.4963 - r_square: 0.7267 - val_loss: 4276.9380 - val_r_square: 0.5019\n",
      "Epoch 213/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2911.5166 - r_square: 0.7168 - val_loss: 4153.1577 - val_r_square: 0.5163\n",
      "Epoch 214/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 3049.9143 - r_square: 0.7034 - val_loss: 3867.5146 - val_r_square: 0.5496\n",
      "Epoch 215/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2853.0210 - r_square: 0.7225 - val_loss: 4168.5229 - val_r_square: 0.5145\n",
      "Epoch 216/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3090.8689 - r_square: 0.6994 - val_loss: 4251.7793 - val_r_square: 0.5048\n",
      "Epoch 217/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3078.3857 - r_square: 0.7006 - val_loss: 3748.1545 - val_r_square: 0.5635\n",
      "Epoch 218/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3070.7422 - r_square: 0.7014 - val_loss: 4388.1777 - val_r_square: 0.4889\n",
      "Epoch 219/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 3157.9197 - r_square: 0.6929 - val_loss: 4093.6987 - val_r_square: 0.5232\n",
      "Epoch 220/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 2914.2869 - r_square: 0.7166 - val_loss: 4026.7444 - val_r_square: 0.5310\n",
      "Epoch 221/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2849.3870 - r_square: 0.7229 - val_loss: 4265.8516 - val_r_square: 0.5032\n",
      "Epoch 222/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2728.8533 - r_square: 0.7346 - val_loss: 6339.6587 - val_r_square: 0.2617\n",
      "Epoch 223/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 3424.7312 - r_square: 0.6669 - val_loss: 5192.3862 - val_r_square: 0.3953\n",
      "Epoch 224/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3113.9934 - r_square: 0.6971 - val_loss: 4528.4658 - val_r_square: 0.4726\n",
      "Epoch 225/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3000.4360 - r_square: 0.7082 - val_loss: 3796.1665 - val_r_square: 0.5579\n",
      "Epoch 226/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3087.1162 - r_square: 0.6998 - val_loss: 4023.4524 - val_r_square: 0.5314\n",
      "Epoch 227/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2969.5542 - r_square: 0.7112 - val_loss: 4444.1807 - val_r_square: 0.4824\n",
      "Epoch 228/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2932.3699 - r_square: 0.7148 - val_loss: 3853.9990 - val_r_square: 0.5512\n",
      "Epoch 229/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2748.4829 - r_square: 0.7327 - val_loss: 3955.5911 - val_r_square: 0.5393\n",
      "Epoch 230/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2646.6997 - r_square: 0.7426 - val_loss: 3902.2710 - val_r_square: 0.5455\n",
      "Epoch 231/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2644.6521 - r_square: 0.7428 - val_loss: 4586.7856 - val_r_square: 0.4658\n",
      "Epoch 232/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2810.4331 - r_square: 0.7267 - val_loss: 4685.0801 - val_r_square: 0.4544\n",
      "Epoch 233/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2764.4390 - r_square: 0.7311 - val_loss: 4320.6763 - val_r_square: 0.4968\n",
      "Epoch 234/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 3033.5576 - r_square: 0.7050 - val_loss: 3989.0732 - val_r_square: 0.5354\n",
      "Epoch 235/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2707.4045 - r_square: 0.7367 - val_loss: 3682.6741 - val_r_square: 0.5711\n",
      "Epoch 236/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2849.2842 - r_square: 0.7229 - val_loss: 4443.0757 - val_r_square: 0.4826\n",
      "Epoch 237/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2799.1636 - r_square: 0.7278 - val_loss: 4424.4570 - val_r_square: 0.4847\n",
      "Epoch 238/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2629.3335 - r_square: 0.7443 - val_loss: 4705.2217 - val_r_square: 0.4520\n",
      "Epoch 239/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2609.5229 - r_square: 0.7462 - val_loss: 5229.1123 - val_r_square: 0.3910\n",
      "Epoch 240/400\n",
      "8/8 [==============================] - 0s 15ms/step - loss: 2930.3882 - r_square: 0.7150 - val_loss: 3964.0017 - val_r_square: 0.5383\n",
      "Epoch 241/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2686.3499 - r_square: 0.7387 - val_loss: 3928.0818 - val_r_square: 0.5425\n",
      "Epoch 242/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2673.5396 - r_square: 0.7400 - val_loss: 4212.5127 - val_r_square: 0.5094\n",
      "Epoch 243/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2727.6802 - r_square: 0.7347 - val_loss: 3944.5103 - val_r_square: 0.5406\n",
      "Epoch 244/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 2833.8179 - r_square: 0.7244 - val_loss: 4136.3774 - val_r_square: 0.5183\n",
      "Epoch 245/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 2696.5007 - r_square: 0.7378 - val_loss: 4690.5122 - val_r_square: 0.4537\n",
      "Epoch 246/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2779.9258 - r_square: 0.7296 - val_loss: 6431.8037 - val_r_square: 0.2509\n",
      "Epoch 247/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 3233.4731 - r_square: 0.6855 - val_loss: 4541.5298 - val_r_square: 0.4711\n",
      "Epoch 248/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 2772.6162 - r_square: 0.7303 - val_loss: 4204.2446 - val_r_square: 0.5104\n",
      "Epoch 249/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2483.1306 - r_square: 0.7585 - val_loss: 3908.1753 - val_r_square: 0.5448\n",
      "Epoch 250/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2523.7600 - r_square: 0.7545 - val_loss: 4418.2715 - val_r_square: 0.4854\n",
      "Epoch 251/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 2612.5723 - r_square: 0.7459 - val_loss: 4732.8926 - val_r_square: 0.4488\n",
      "Epoch 252/400\n",
      "8/8 [==============================] - 0s 13ms/step - loss: 2551.6641 - r_square: 0.7518 - val_loss: 4736.1011 - val_r_square: 0.4484\n",
      "Epoch 253/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2647.5378 - r_square: 0.7425 - val_loss: 4158.9600 - val_r_square: 0.5156\n",
      "Epoch 254/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2568.8130 - r_square: 0.7502 - val_loss: 4713.9399 - val_r_square: 0.4510\n",
      "Epoch 255/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2581.9810 - r_square: 0.7489 - val_loss: 3977.6895 - val_r_square: 0.5368\n",
      "Epoch 256/400\n",
      "8/8 [==============================] - 0s 14ms/step - loss: 2729.9800 - r_square: 0.7345 - val_loss: 4257.3374 - val_r_square: 0.5042\n",
      "Epoch 257/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2663.2129 - r_square: 0.7410 - val_loss: 3788.3364 - val_r_square: 0.5588\n",
      "Epoch 258/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2638.5837 - r_square: 0.7434 - val_loss: 4475.2378 - val_r_square: 0.4788\n",
      "Epoch 259/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 2508.8523 - r_square: 0.7560 - val_loss: 4577.8330 - val_r_square: 0.4669\n",
      "Epoch 260/400\n",
      "8/8 [==============================] - 0s 12ms/step - loss: 2437.7610 - r_square: 0.7629 - val_loss: 5958.2119 - val_r_square: 0.3061\n",
      "Epoch 261/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 3030.2444 - r_square: 0.7053 - val_loss: 5012.8525 - val_r_square: 0.4162\n",
      "Epoch 262/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2500.6689 - r_square: 0.7568 - val_loss: 4274.5107 - val_r_square: 0.5022\n",
      "Epoch 263/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2758.6804 - r_square: 0.7317 - val_loss: 3943.9226 - val_r_square: 0.5407\n",
      "Epoch 264/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2771.7722 - r_square: 0.7304 - val_loss: 3961.6931 - val_r_square: 0.5386\n",
      "Epoch 265/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2711.6565 - r_square: 0.7363 - val_loss: 4858.9038 - val_r_square: 0.4341\n",
      "Epoch 266/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2707.6028 - r_square: 0.7367 - val_loss: 5506.3940 - val_r_square: 0.3587\n",
      "Epoch 267/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2529.5906 - r_square: 0.7540 - val_loss: 4979.8135 - val_r_square: 0.4200\n",
      "Epoch 268/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2471.0769 - r_square: 0.7597 - val_loss: 4373.4238 - val_r_square: 0.4907\n",
      "Epoch 269/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2421.1147 - r_square: 0.7645 - val_loss: 3787.8035 - val_r_square: 0.5589\n",
      "Epoch 270/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2481.6091 - r_square: 0.7586 - val_loss: 3942.4038 - val_r_square: 0.5409\n",
      "Epoch 271/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2683.0615 - r_square: 0.7391 - val_loss: 6645.7119 - val_r_square: 0.2260\n",
      "Epoch 272/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 3263.5413 - r_square: 0.6826 - val_loss: 5344.7437 - val_r_square: 0.3775\n",
      "Epoch 273/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2689.3630 - r_square: 0.7384 - val_loss: 4018.8013 - val_r_square: 0.5320\n",
      "Epoch 274/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2510.3872 - r_square: 0.7559 - val_loss: 4400.9287 - val_r_square: 0.4875\n",
      "Epoch 275/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2397.6758 - r_square: 0.7668 - val_loss: 3742.4060 - val_r_square: 0.5642\n",
      "Epoch 276/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2403.5923 - r_square: 0.7662 - val_loss: 3704.5623 - val_r_square: 0.5686\n",
      "Epoch 277/400\n",
      "8/8 [==============================] - 1s 85ms/step - loss: 2479.7185 - r_square: 0.7588 - val_loss: 4871.7695 - val_r_square: 0.4326\n",
      "Epoch 278/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 2728.0696 - r_square: 0.7347 - val_loss: 5445.3457 - val_r_square: 0.3658\n",
      "Epoch 279/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2448.1912 - r_square: 0.7619 - val_loss: 4380.8901 - val_r_square: 0.4898\n",
      "Epoch 280/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2352.2800 - r_square: 0.7712 - val_loss: 4036.3892 - val_r_square: 0.5299\n",
      "Epoch 281/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2354.6843 - r_square: 0.7710 - val_loss: 3675.2314 - val_r_square: 0.5720\n",
      "Epoch 282/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2377.1416 - r_square: 0.7688 - val_loss: 4035.3909 - val_r_square: 0.5300\n",
      "Epoch 283/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2408.2014 - r_square: 0.7658 - val_loss: 4172.7505 - val_r_square: 0.5140\n",
      "Epoch 284/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2666.1753 - r_square: 0.7407 - val_loss: 5772.9609 - val_r_square: 0.3277\n",
      "Epoch 285/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2789.6326 - r_square: 0.7287 - val_loss: 4221.2139 - val_r_square: 0.5084\n",
      "Epoch 286/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2476.1560 - r_square: 0.7592 - val_loss: 3862.5596 - val_r_square: 0.5502\n",
      "Epoch 287/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2506.4189 - r_square: 0.7562 - val_loss: 3698.5115 - val_r_square: 0.5693\n",
      "Epoch 288/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2204.6797 - r_square: 0.7856 - val_loss: 3786.4011 - val_r_square: 0.5590\n",
      "Epoch 289/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2309.9255 - r_square: 0.7753 - val_loss: 3748.4705 - val_r_square: 0.5634\n",
      "Epoch 290/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2220.6421 - r_square: 0.7840 - val_loss: 4101.3838 - val_r_square: 0.5223\n",
      "Epoch 291/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2175.2493 - r_square: 0.7884 - val_loss: 3962.3813 - val_r_square: 0.5385\n",
      "Epoch 292/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2278.6392 - r_square: 0.7784 - val_loss: 4201.2915 - val_r_square: 0.5107\n",
      "Epoch 293/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2431.5051 - r_square: 0.7635 - val_loss: 5374.3091 - val_r_square: 0.3741\n",
      "Epoch 294/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2380.5007 - r_square: 0.7685 - val_loss: 4766.4082 - val_r_square: 0.4449\n",
      "Epoch 295/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2211.8464 - r_square: 0.7849 - val_loss: 4017.6113 - val_r_square: 0.5321\n",
      "Epoch 296/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2185.4563 - r_square: 0.7875 - val_loss: 4679.2690 - val_r_square: 0.4550\n",
      "Epoch 297/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2385.1577 - r_square: 0.7680 - val_loss: 4035.1682 - val_r_square: 0.5301\n",
      "Epoch 298/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2146.6682 - r_square: 0.7912 - val_loss: 3976.0486 - val_r_square: 0.5369\n",
      "Epoch 299/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 2258.0122 - r_square: 0.7804 - val_loss: 4157.1792 - val_r_square: 0.5158\n",
      "Epoch 300/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2422.1589 - r_square: 0.7644 - val_loss: 4655.5498 - val_r_square: 0.4578\n",
      "Epoch 301/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2215.7263 - r_square: 0.7845 - val_loss: 4615.0986 - val_r_square: 0.4625\n",
      "Epoch 302/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2240.9297 - r_square: 0.7821 - val_loss: 4166.7319 - val_r_square: 0.5147\n",
      "Epoch 303/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2209.3638 - r_square: 0.7851 - val_loss: 4102.3418 - val_r_square: 0.5222\n",
      "Epoch 304/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2476.9592 - r_square: 0.7591 - val_loss: 4175.7334 - val_r_square: 0.5137\n",
      "Epoch 305/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2252.9365 - r_square: 0.7809 - val_loss: 4352.9214 - val_r_square: 0.4931\n",
      "Epoch 306/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2149.4597 - r_square: 0.7910 - val_loss: 4953.1138 - val_r_square: 0.4232\n",
      "Epoch 307/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2193.8210 - r_square: 0.7866 - val_loss: 6224.4683 - val_r_square: 0.2751\n",
      "Epoch 308/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2303.9268 - r_square: 0.7759 - val_loss: 4857.2695 - val_r_square: 0.4343\n",
      "Epoch 309/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2244.2773 - r_square: 0.7817 - val_loss: 4277.1592 - val_r_square: 0.5019\n",
      "Epoch 310/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2088.7017 - r_square: 0.7969 - val_loss: 4333.5806 - val_r_square: 0.4953\n",
      "Epoch 311/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2047.9977 - r_square: 0.8008 - val_loss: 4035.2100 - val_r_square: 0.5301\n",
      "Epoch 312/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2150.0518 - r_square: 0.7909 - val_loss: 4349.1206 - val_r_square: 0.4935\n",
      "Epoch 313/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2221.3254 - r_square: 0.7840 - val_loss: 5417.0186 - val_r_square: 0.3691\n",
      "Epoch 314/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2313.7129 - r_square: 0.7750 - val_loss: 4843.5952 - val_r_square: 0.4359\n",
      "Epoch 315/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2290.5508 - r_square: 0.7772 - val_loss: 3944.0618 - val_r_square: 0.5407\n",
      "Epoch 316/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1953.3275 - r_square: 0.8100 - val_loss: 4526.4736 - val_r_square: 0.4728\n",
      "Epoch 317/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2050.2976 - r_square: 0.8006 - val_loss: 5142.5864 - val_r_square: 0.4011\n",
      "Epoch 318/400\n",
      "8/8 [==============================] - 0s 10ms/step - loss: 2272.4202 - r_square: 0.7790 - val_loss: 3901.1096 - val_r_square: 0.5457\n",
      "Epoch 319/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2255.4031 - r_square: 0.7806 - val_loss: 3989.1670 - val_r_square: 0.5354\n",
      "Epoch 320/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1974.4568 - r_square: 0.8080 - val_loss: 4208.4570 - val_r_square: 0.5099\n",
      "Epoch 321/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2010.8335 - r_square: 0.8044 - val_loss: 5533.0195 - val_r_square: 0.3556\n",
      "Epoch 322/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2368.2485 - r_square: 0.7697 - val_loss: 4656.6064 - val_r_square: 0.4577\n",
      "Epoch 323/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2278.5288 - r_square: 0.7784 - val_loss: 4075.6333 - val_r_square: 0.5253\n",
      "Epoch 324/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2036.0803 - r_square: 0.8020 - val_loss: 4489.1987 - val_r_square: 0.4772\n",
      "Epoch 325/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2118.5791 - r_square: 0.7940 - val_loss: 6057.0308 - val_r_square: 0.2946\n",
      "Epoch 326/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2164.0378 - r_square: 0.7895 - val_loss: 4271.2231 - val_r_square: 0.5026\n",
      "Epoch 327/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1928.7799 - r_square: 0.8124 - val_loss: 3799.3259 - val_r_square: 0.5575\n",
      "Epoch 328/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1914.4233 - r_square: 0.8138 - val_loss: 4656.7676 - val_r_square: 0.4577\n",
      "Epoch 329/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2033.6871 - r_square: 0.8022 - val_loss: 4271.6396 - val_r_square: 0.5025\n",
      "Epoch 330/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1892.4518 - r_square: 0.8159 - val_loss: 4032.0508 - val_r_square: 0.5304\n",
      "Epoch 331/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1827.9310 - r_square: 0.8222 - val_loss: 4259.2261 - val_r_square: 0.5040\n",
      "Epoch 332/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1778.4120 - r_square: 0.8270 - val_loss: 4556.4707 - val_r_square: 0.4693\n",
      "Epoch 333/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1770.8942 - r_square: 0.8278 - val_loss: 4155.9316 - val_r_square: 0.5160\n",
      "Epoch 334/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1927.2002 - r_square: 0.8126 - val_loss: 3994.0654 - val_r_square: 0.5348\n",
      "Epoch 335/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2007.8400 - r_square: 0.8047 - val_loss: 5784.0054 - val_r_square: 0.3264\n",
      "Epoch 336/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2438.1487 - r_square: 0.7629 - val_loss: 4763.3335 - val_r_square: 0.4453\n",
      "Epoch 337/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2494.3118 - r_square: 0.7574 - val_loss: 3886.0410 - val_r_square: 0.5474\n",
      "Epoch 338/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2160.3660 - r_square: 0.7899 - val_loss: 4569.7109 - val_r_square: 0.4678\n",
      "Epoch 339/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1791.9767 - r_square: 0.8257 - val_loss: 3947.9673 - val_r_square: 0.5402\n",
      "Epoch 340/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1956.4032 - r_square: 0.8097 - val_loss: 4391.5825 - val_r_square: 0.4886\n",
      "Epoch 341/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2089.8289 - r_square: 0.7968 - val_loss: 3906.6233 - val_r_square: 0.5450\n",
      "Epoch 342/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2255.1372 - r_square: 0.7807 - val_loss: 3925.8237 - val_r_square: 0.5428\n",
      "Epoch 343/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2064.6877 - r_square: 0.7992 - val_loss: 5151.3906 - val_r_square: 0.4001\n",
      "Epoch 344/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2103.4673 - r_square: 0.7954 - val_loss: 4029.4978 - val_r_square: 0.5307\n",
      "Epoch 345/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1866.2823 - r_square: 0.8185 - val_loss: 4083.7434 - val_r_square: 0.5244\n",
      "Epoch 346/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1726.1958 - r_square: 0.8321 - val_loss: 4114.0972 - val_r_square: 0.5209\n",
      "Epoch 347/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1641.8223 - r_square: 0.8403 - val_loss: 4214.4585 - val_r_square: 0.5092\n",
      "Epoch 348/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1728.4325 - r_square: 0.8319 - val_loss: 4924.4497 - val_r_square: 0.4265\n",
      "Epoch 349/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1959.9791 - r_square: 0.8094 - val_loss: 4112.3896 - val_r_square: 0.5211\n",
      "Epoch 350/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1886.5055 - r_square: 0.8165 - val_loss: 4795.4292 - val_r_square: 0.4415\n",
      "Epoch 351/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1796.4458 - r_square: 0.8253 - val_loss: 4753.7129 - val_r_square: 0.4464\n",
      "Epoch 352/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1834.3857 - r_square: 0.8216 - val_loss: 3953.3145 - val_r_square: 0.5396\n",
      "Epoch 353/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1929.5170 - r_square: 0.8123 - val_loss: 5077.4580 - val_r_square: 0.4087\n",
      "Epoch 354/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1918.5919 - r_square: 0.8134 - val_loss: 5429.0635 - val_r_square: 0.3677\n",
      "Epoch 355/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1884.3313 - r_square: 0.8167 - val_loss: 4219.2363 - val_r_square: 0.5086\n",
      "Epoch 356/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1954.8550 - r_square: 0.8099 - val_loss: 3890.6475 - val_r_square: 0.5469\n",
      "Epoch 357/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1958.3250 - r_square: 0.8095 - val_loss: 4290.6279 - val_r_square: 0.5003\n",
      "Epoch 358/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1975.5793 - r_square: 0.8079 - val_loss: 5592.1582 - val_r_square: 0.3487\n",
      "Epoch 359/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2005.0227 - r_square: 0.8050 - val_loss: 4319.0044 - val_r_square: 0.4970\n",
      "Epoch 360/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 2029.7659 - r_square: 0.8026 - val_loss: 4341.8452 - val_r_square: 0.4943\n",
      "Epoch 361/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2312.8916 - r_square: 0.7751 - val_loss: 4783.0522 - val_r_square: 0.4430\n",
      "Epoch 362/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1878.9742 - r_square: 0.8173 - val_loss: 4369.2163 - val_r_square: 0.4912\n",
      "Epoch 363/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1629.4875 - r_square: 0.8415 - val_loss: 4041.1560 - val_r_square: 0.5294\n",
      "Epoch 364/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1513.1038 - r_square: 0.8528 - val_loss: 4353.5103 - val_r_square: 0.4930\n",
      "Epoch 365/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1539.2609 - r_square: 0.8503 - val_loss: 4688.6255 - val_r_square: 0.4540\n",
      "Epoch 366/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1731.1447 - r_square: 0.8316 - val_loss: 4411.2393 - val_r_square: 0.4863\n",
      "Epoch 367/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1690.2788 - r_square: 0.8356 - val_loss: 4467.4673 - val_r_square: 0.4797\n",
      "Epoch 368/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1832.3890 - r_square: 0.8218 - val_loss: 5535.4351 - val_r_square: 0.3553\n",
      "Epoch 369/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1892.2416 - r_square: 0.8160 - val_loss: 4446.3691 - val_r_square: 0.4822\n",
      "Epoch 370/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1729.4083 - r_square: 0.8318 - val_loss: 4123.7207 - val_r_square: 0.5197\n",
      "Epoch 371/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1632.5691 - r_square: 0.8412 - val_loss: 4593.5796 - val_r_square: 0.4650\n",
      "Epoch 372/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1583.5597 - r_square: 0.8460 - val_loss: 5963.4180 - val_r_square: 0.3055\n",
      "Epoch 373/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 2041.2072 - r_square: 0.8015 - val_loss: 4515.3721 - val_r_square: 0.4741\n",
      "Epoch 374/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1822.2708 - r_square: 0.8228 - val_loss: 4745.9214 - val_r_square: 0.4473\n",
      "Epoch 375/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1718.8365 - r_square: 0.8328 - val_loss: 5182.2227 - val_r_square: 0.3965\n",
      "Epoch 376/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1676.8750 - r_square: 0.8369 - val_loss: 3883.2124 - val_r_square: 0.5478\n",
      "Epoch 377/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1609.9799 - r_square: 0.8434 - val_loss: 3856.7852 - val_r_square: 0.5508\n",
      "Epoch 378/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1580.0532 - r_square: 0.8463 - val_loss: 5062.5518 - val_r_square: 0.4104\n",
      "Epoch 379/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1612.0007 - r_square: 0.8432 - val_loss: 4330.9292 - val_r_square: 0.4956\n",
      "Epoch 380/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1673.8712 - r_square: 0.8372 - val_loss: 4127.5776 - val_r_square: 0.5193\n",
      "Epoch 381/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1581.7662 - r_square: 0.8462 - val_loss: 3997.6360 - val_r_square: 0.5344\n",
      "Epoch 382/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1548.6899 - r_square: 0.8494 - val_loss: 4344.7373 - val_r_square: 0.4940\n",
      "Epoch 383/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1417.1945 - r_square: 0.8622 - val_loss: 4210.0737 - val_r_square: 0.5097\n",
      "Epoch 384/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1541.5193 - r_square: 0.8501 - val_loss: 4191.8296 - val_r_square: 0.5118\n",
      "Epoch 385/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1540.6052 - r_square: 0.8502 - val_loss: 4023.0640 - val_r_square: 0.5315\n",
      "Epoch 386/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1668.2035 - r_square: 0.8378 - val_loss: 4314.8301 - val_r_square: 0.4975\n",
      "Epoch 387/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1477.7637 - r_square: 0.8563 - val_loss: 4362.1890 - val_r_square: 0.4920\n",
      "Epoch 388/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1491.7456 - r_square: 0.8549 - val_loss: 4147.5576 - val_r_square: 0.5170\n",
      "Epoch 389/400\n",
      "8/8 [==============================] - 0s 11ms/step - loss: 1394.9163 - r_square: 0.8643 - val_loss: 4317.7012 - val_r_square: 0.4972\n",
      "Epoch 390/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1364.2988 - r_square: 0.8673 - val_loss: 3902.9089 - val_r_square: 0.5455\n",
      "Epoch 391/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1416.6377 - r_square: 0.8622 - val_loss: 4171.0337 - val_r_square: 0.5142\n",
      "Epoch 392/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1713.9113 - r_square: 0.8333 - val_loss: 4760.5317 - val_r_square: 0.4456\n",
      "Epoch 393/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2029.8844 - r_square: 0.8026 - val_loss: 5056.8823 - val_r_square: 0.4111\n",
      "Epoch 394/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1658.3805 - r_square: 0.8387 - val_loss: 4827.7397 - val_r_square: 0.4378\n",
      "Epoch 395/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1669.4205 - r_square: 0.8376 - val_loss: 3902.6387 - val_r_square: 0.5455\n",
      "Epoch 396/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 1572.0282 - r_square: 0.8471 - val_loss: 4419.0571 - val_r_square: 0.4854\n",
      "Epoch 397/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1358.2518 - r_square: 0.8679 - val_loss: 5308.0991 - val_r_square: 0.3818\n",
      "Epoch 398/400\n",
      "8/8 [==============================] - 0s 7ms/step - loss: 1866.2234 - r_square: 0.8185 - val_loss: 3954.3040 - val_r_square: 0.5395\n",
      "Epoch 399/400\n",
      "8/8 [==============================] - 0s 9ms/step - loss: 1800.3506 - r_square: 0.8249 - val_loss: 3883.6206 - val_r_square: 0.5477\n",
      "Epoch 400/400\n",
      "8/8 [==============================] - 0s 8ms/step - loss: 2126.1887 - r_square: 0.7932 - val_loss: 5524.1455 - val_r_square: 0.3567\n"
     ]
    }
   ],
   "source": [
    "history_r1 = model_relu1.fit(X_train, y_train, epochs=400, batch_size=128, validation_data=(X_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "id": "d9fe104c-bbea-4b25-bd4f-35e2e842e556",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 0s 2ms/step\n",
      "44.67220296893047\n"
     ]
    }
   ],
   "source": [
    "y_pred_relu_tr = model_relu1.predict(X_train)\n",
    "relu1_rmse_train = np.sqrt(mean_squared_error(y_train, y_pred_relu_tr))\n",
    "print(relu1_rmse_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 285,
   "id": "d2d383ab-3f25-4dfc-9f27-53db1985013d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9/9 [==============================] - 0s 3ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "287"
      ]
     },
     "execution_count": 285,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions_relu = model_relu1.predict(X_test)\n",
    "len(predictions_relu)\n",
    "#predictions_l"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 286,
   "id": "89f784d6-bb5c-410a-bb30-93fd9e9ddd4a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "72.6125570903605\n"
     ]
    }
   ],
   "source": [
    "relu1_rmse = np.sqrt(mean_squared_error(y_test, predictions_relu))\n",
    "print(relu1_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 413,
   "id": "99dbe35f-45f9-4648-bdc1-f3ed42e23c7f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots()\n",
    "fig.set_figheight(4)\n",
    "fig.set_figwidth(15)\n",
    "\n",
    "axs.plot(predictions_relu,color='red', label='Predicted')\n",
    "axs.plot(y_test,color='blue', label='Actual')\n",
    "#axs.scatter(outliers.index,outliers, color='green', linewidth=5.0, label='Anomalies')\n",
    "plt.xlabel('Timestamp')\n",
    "plt.ylabel('Scaled number of passengers')\n",
    "plt.legend(loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 287,
   "id": "82dcd7b4-a342-4980-ab6d-af314536a8ff",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 4ms/step - loss: 5272.5835 - r_square: 0.4434\n"
     ]
    }
   ],
   "source": [
    "results = model_relu1.evaluate(X_test, y_test, batch_size=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 320,
   "id": "d147114a-210e-430c-9dc8-0bb7f98b1733",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib\n",
    "\n",
    "error = predictions_relu - y_test\n",
    "plt.hist(error, bins=25)\n",
    "plt.xlabel('Prediction Error')\n",
    "_ = plt.ylabel('Count')\n",
    "\n",
    "fig = matplotlib.pyplot.gcf()\n",
    "fig.set_size_inches(10,10)\n",
    "fig.savefig('pred_error_relu1.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 321,
   "id": "84f4dcaa-5851-4b40-a0e0-7c44e14630cd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history_r1)\n",
    "plt.savefig('PlotLoss_ModelRelu1.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b10c16af-604c-4e5a-8be7-9dd675773f54",
   "metadata": {},
   "source": [
    "## LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f6d1cdbd-2b75-45e8-9402-6bab062f14d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras import layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "id": "009cef37-a66f-4054-aa38-2a0d3cb036d3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(916, 30, 1)"
      ]
     },
     "execution_count": 290,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_reshaped = X_train.reshape(X_train.shape[0], X_train.shape[1], 1) \n",
    "X_train_reshaped.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "63535f82-9227-4024-81d5-3b657da785c2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "id": "efa25b67-a15e-4fbf-beb5-fa482a51ddd6",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''seq_length = X_train.shape[0]\n",
    "num_features = X_train.shape[1]\n",
    "\n",
    "#input_shape = (seq_length, num_features)\n",
    "\n",
    "model_lstm = tf.keras.Sequential([\n",
    "    layers.LSTM(64, input_shape=(num_features, 1)),\n",
    "    layers.Dense(1, activation='linear')\n",
    "])\n",
    "model_lstm.compile(loss='mse', optimizer=tf.keras.optimizers.Adam(learning_rate=0.1), metrics=[tfa.metrics.RSquare()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "id": "a387839a-7938-44e3-b116-f74a2722667f",
   "metadata": {},
   "outputs": [],
   "source": [
    "#history_lstm = model_lstm.fit(X_train_reshaped, y_train, batch_size=128, epochs=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 294,
   "id": "d3089919-2f60-481d-aee0-dbfd226aa54b",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''seq_length = X_train.shape[0]\n",
    "num_features = X_train.shape[1]\n",
    "input_shape = (seq_length, num_features)\n",
    "\n",
    "model_lstm = tf.keras.Sequential([\n",
    "    layers.LSTM(64, input_shape=input_shape),\n",
    "    layers.Dense(1, activation='linear')\n",
    "])\n",
    "model_lstm.compile(loss='mse', optimizer='adam', metrics=[tfa.metrics.RSquare()])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5b67ec98-d643-44ec-8193-6fe1e4cea7b8",
   "metadata": {},
   "source": [
    "## Changing the shape of data to 3D to accomodate 8 time points for 5 features each"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 300,
   "id": "a09bcc38-1a9c-48da-8c17-cc4ecf81db0c",
   "metadata": {},
   "outputs": [],
   "source": [
    "time_steps = 6\n",
    "num_features = 5\n",
    "samples = df_X.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 301,
   "id": "2d7f4eae-70f8-42ee-8ec0-f384bb74b654",
   "metadata": {},
   "outputs": [],
   "source": [
    "#X_train_rs = X_train[:samples*time_steps]\n",
    "#X_train_rs = np.reshape(X_train, (samples, time_steps, num_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 302,
   "id": "57b1ff43-a351-4204-93d9-73e7ac89b824",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.453394</td>\n",
       "      <td>-1.114558</td>\n",
       "      <td>1.315852</td>\n",
       "      <td>-0.647019</td>\n",
       "      <td>-1.156884</td>\n",
       "      <td>-0.915120</td>\n",
       "      <td>-0.467190</td>\n",
       "      <td>-0.971554</td>\n",
       "      <td>1.181514</td>\n",
       "      <td>-0.324311</td>\n",
       "      <td>...</td>\n",
       "      <td>1.316829</td>\n",
       "      <td>-0.915700</td>\n",
       "      <td>-0.391924</td>\n",
       "      <td>-0.361832</td>\n",
       "      <td>0.134009</td>\n",
       "      <td>-0.815805</td>\n",
       "      <td>1.160813</td>\n",
       "      <td>-0.763477</td>\n",
       "      <td>-0.509573</td>\n",
       "      <td>-1.119741</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.146211</td>\n",
       "      <td>-1.041571</td>\n",
       "      <td>1.366997</td>\n",
       "      <td>-1.206831</td>\n",
       "      <td>-1.038537</td>\n",
       "      <td>0.615373</td>\n",
       "      <td>-0.259258</td>\n",
       "      <td>-0.874142</td>\n",
       "      <td>1.253709</td>\n",
       "      <td>-0.990028</td>\n",
       "      <td>...</td>\n",
       "      <td>1.232367</td>\n",
       "      <td>-1.111031</td>\n",
       "      <td>-0.026364</td>\n",
       "      <td>0.823203</td>\n",
       "      <td>0.126408</td>\n",
       "      <td>-0.793194</td>\n",
       "      <td>1.180831</td>\n",
       "      <td>-1.129151</td>\n",
       "      <td>-0.223885</td>\n",
       "      <td>-0.168157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.012917</td>\n",
       "      <td>-0.814373</td>\n",
       "      <td>1.235048</td>\n",
       "      <td>-0.983579</td>\n",
       "      <td>-1.403975</td>\n",
       "      <td>-0.890483</td>\n",
       "      <td>-0.210613</td>\n",
       "      <td>-0.803763</td>\n",
       "      <td>1.478827</td>\n",
       "      <td>-1.278609</td>\n",
       "      <td>...</td>\n",
       "      <td>1.749245</td>\n",
       "      <td>-1.268517</td>\n",
       "      <td>-0.618770</td>\n",
       "      <td>0.010361</td>\n",
       "      <td>0.182520</td>\n",
       "      <td>-0.791873</td>\n",
       "      <td>1.331832</td>\n",
       "      <td>-1.321442</td>\n",
       "      <td>-0.845928</td>\n",
       "      <td>-0.989878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.128726</td>\n",
       "      <td>-0.863284</td>\n",
       "      <td>1.344185</td>\n",
       "      <td>-1.207989</td>\n",
       "      <td>-1.941000</td>\n",
       "      <td>-0.965864</td>\n",
       "      <td>-0.429344</td>\n",
       "      <td>-0.890786</td>\n",
       "      <td>1.426227</td>\n",
       "      <td>-1.115059</td>\n",
       "      <td>...</td>\n",
       "      <td>1.462318</td>\n",
       "      <td>-0.814971</td>\n",
       "      <td>-0.675256</td>\n",
       "      <td>-1.024408</td>\n",
       "      <td>0.025213</td>\n",
       "      <td>-0.849368</td>\n",
       "      <td>1.278458</td>\n",
       "      <td>-1.006951</td>\n",
       "      <td>-0.777635</td>\n",
       "      <td>-1.208612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.455706</td>\n",
       "      <td>-0.833586</td>\n",
       "      <td>1.301384</td>\n",
       "      <td>-1.550175</td>\n",
       "      <td>-1.638518</td>\n",
       "      <td>0.111595</td>\n",
       "      <td>-0.054254</td>\n",
       "      <td>-0.817844</td>\n",
       "      <td>1.305953</td>\n",
       "      <td>-0.641529</td>\n",
       "      <td>...</td>\n",
       "      <td>1.280126</td>\n",
       "      <td>-0.331646</td>\n",
       "      <td>-0.157677</td>\n",
       "      <td>-1.078575</td>\n",
       "      <td>0.226101</td>\n",
       "      <td>-0.797131</td>\n",
       "      <td>1.197110</td>\n",
       "      <td>-0.448046</td>\n",
       "      <td>-0.662992</td>\n",
       "      <td>-1.835126</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1427</th>\n",
       "      <td>1.292526</td>\n",
       "      <td>-0.694296</td>\n",
       "      <td>0.085300</td>\n",
       "      <td>0.070252</td>\n",
       "      <td>1.150916</td>\n",
       "      <td>-0.773195</td>\n",
       "      <td>1.454364</td>\n",
       "      <td>-0.713077</td>\n",
       "      <td>-0.020875</td>\n",
       "      <td>-0.023632</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.188384</td>\n",
       "      <td>-0.389774</td>\n",
       "      <td>1.314533</td>\n",
       "      <td>0.982503</td>\n",
       "      <td>1.409141</td>\n",
       "      <td>-0.702248</td>\n",
       "      <td>0.122504</td>\n",
       "      <td>0.171835</td>\n",
       "      <td>1.397982</td>\n",
       "      <td>-0.572173</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1428</th>\n",
       "      <td>1.152732</td>\n",
       "      <td>-0.746422</td>\n",
       "      <td>0.175350</td>\n",
       "      <td>0.059034</td>\n",
       "      <td>0.571081</td>\n",
       "      <td>-0.224036</td>\n",
       "      <td>1.298967</td>\n",
       "      <td>-0.778784</td>\n",
       "      <td>0.110101</td>\n",
       "      <td>-0.225567</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.043797</td>\n",
       "      <td>-0.480007</td>\n",
       "      <td>1.063266</td>\n",
       "      <td>1.276821</td>\n",
       "      <td>1.208258</td>\n",
       "      <td>-0.767089</td>\n",
       "      <td>0.243979</td>\n",
       "      <td>-0.177111</td>\n",
       "      <td>1.363294</td>\n",
       "      <td>-0.472232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1429</th>\n",
       "      <td>1.290085</td>\n",
       "      <td>-0.660370</td>\n",
       "      <td>0.091188</td>\n",
       "      <td>0.045040</td>\n",
       "      <td>0.259596</td>\n",
       "      <td>0.034854</td>\n",
       "      <td>1.440994</td>\n",
       "      <td>-0.666851</td>\n",
       "      <td>-0.026565</td>\n",
       "      <td>-0.114877</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.196349</td>\n",
       "      <td>-0.400987</td>\n",
       "      <td>0.680580</td>\n",
       "      <td>0.313022</td>\n",
       "      <td>1.323714</td>\n",
       "      <td>-0.714102</td>\n",
       "      <td>0.138415</td>\n",
       "      <td>0.010560</td>\n",
       "      <td>0.988390</td>\n",
       "      <td>0.073366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1430</th>\n",
       "      <td>1.273268</td>\n",
       "      <td>-0.658241</td>\n",
       "      <td>0.131593</td>\n",
       "      <td>-0.017340</td>\n",
       "      <td>0.815052</td>\n",
       "      <td>0.245985</td>\n",
       "      <td>1.362855</td>\n",
       "      <td>-0.671782</td>\n",
       "      <td>0.041397</td>\n",
       "      <td>-0.242208</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.125857</td>\n",
       "      <td>-0.449296</td>\n",
       "      <td>1.021932</td>\n",
       "      <td>1.102812</td>\n",
       "      <td>1.269366</td>\n",
       "      <td>-0.722324</td>\n",
       "      <td>0.187538</td>\n",
       "      <td>-0.145163</td>\n",
       "      <td>1.294288</td>\n",
       "      <td>-0.117775</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1431</th>\n",
       "      <td>0.876400</td>\n",
       "      <td>-0.707958</td>\n",
       "      <td>0.183111</td>\n",
       "      <td>0.114643</td>\n",
       "      <td>0.595369</td>\n",
       "      <td>-0.241087</td>\n",
       "      <td>1.020279</td>\n",
       "      <td>-0.738615</td>\n",
       "      <td>0.041696</td>\n",
       "      <td>0.078013</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.121707</td>\n",
       "      <td>-0.306473</td>\n",
       "      <td>0.935626</td>\n",
       "      <td>-0.319542</td>\n",
       "      <td>0.929498</td>\n",
       "      <td>-0.747676</td>\n",
       "      <td>0.193786</td>\n",
       "      <td>0.223996</td>\n",
       "      <td>1.209726</td>\n",
       "      <td>-0.532616</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1432 rows × 30 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            0         1         2         3         4         5         6   \\\n",
       "0    -0.453394 -1.114558  1.315852 -0.647019 -1.156884 -0.915120 -0.467190   \n",
       "1    -0.146211 -1.041571  1.366997 -1.206831 -1.038537  0.615373 -0.259258   \n",
       "2     0.012917 -0.814373  1.235048 -0.983579 -1.403975 -0.890483 -0.210613   \n",
       "3    -0.128726 -0.863284  1.344185 -1.207989 -1.941000 -0.965864 -0.429344   \n",
       "4     0.455706 -0.833586  1.301384 -1.550175 -1.638518  0.111595 -0.054254   \n",
       "...        ...       ...       ...       ...       ...       ...       ...   \n",
       "1427  1.292526 -0.694296  0.085300  0.070252  1.150916 -0.773195  1.454364   \n",
       "1428  1.152732 -0.746422  0.175350  0.059034  0.571081 -0.224036  1.298967   \n",
       "1429  1.290085 -0.660370  0.091188  0.045040  0.259596  0.034854  1.440994   \n",
       "1430  1.273268 -0.658241  0.131593 -0.017340  0.815052  0.245985  1.362855   \n",
       "1431  0.876400 -0.707958  0.183111  0.114643  0.595369 -0.241087  1.020279   \n",
       "\n",
       "            7         8         9   ...        20        21        22  \\\n",
       "0    -0.971554  1.181514 -0.324311  ...  1.316829 -0.915700 -0.391924   \n",
       "1    -0.874142  1.253709 -0.990028  ...  1.232367 -1.111031 -0.026364   \n",
       "2    -0.803763  1.478827 -1.278609  ...  1.749245 -1.268517 -0.618770   \n",
       "3    -0.890786  1.426227 -1.115059  ...  1.462318 -0.814971 -0.675256   \n",
       "4    -0.817844  1.305953 -0.641529  ...  1.280126 -0.331646 -0.157677   \n",
       "...        ...       ...       ...  ...       ...       ...       ...   \n",
       "1427 -0.713077 -0.020875 -0.023632  ... -0.188384 -0.389774  1.314533   \n",
       "1428 -0.778784  0.110101 -0.225567  ... -0.043797 -0.480007  1.063266   \n",
       "1429 -0.666851 -0.026565 -0.114877  ... -0.196349 -0.400987  0.680580   \n",
       "1430 -0.671782  0.041397 -0.242208  ... -0.125857 -0.449296  1.021932   \n",
       "1431 -0.738615  0.041696  0.078013  ... -0.121707 -0.306473  0.935626   \n",
       "\n",
       "            23        24        25        26        27        28        29  \n",
       "0    -0.361832  0.134009 -0.815805  1.160813 -0.763477 -0.509573 -1.119741  \n",
       "1     0.823203  0.126408 -0.793194  1.180831 -1.129151 -0.223885 -0.168157  \n",
       "2     0.010361  0.182520 -0.791873  1.331832 -1.321442 -0.845928 -0.989878  \n",
       "3    -1.024408  0.025213 -0.849368  1.278458 -1.006951 -0.777635 -1.208612  \n",
       "4    -1.078575  0.226101 -0.797131  1.197110 -0.448046 -0.662992 -1.835126  \n",
       "...        ...       ...       ...       ...       ...       ...       ...  \n",
       "1427  0.982503  1.409141 -0.702248  0.122504  0.171835  1.397982 -0.572173  \n",
       "1428  1.276821  1.208258 -0.767089  0.243979 -0.177111  1.363294 -0.472232  \n",
       "1429  0.313022  1.323714 -0.714102  0.138415  0.010560  0.988390  0.073366  \n",
       "1430  1.102812  1.269366 -0.722324  0.187538 -0.145163  1.294288 -0.117775  \n",
       "1431 -0.319542  0.929498 -0.747676  0.193786  0.223996  1.209726 -0.532616  \n",
       "\n",
       "[1432 rows x 30 columns]"
      ]
     },
     "execution_count": 302,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_X = pd.DataFrame(X_sc)\n",
    "df_X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 303,
   "id": "036669da-2419-4f75-b5db-ecfe8cea70bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_rs = np.zeros((samples, time_steps, num_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 304,
   "id": "e1d1be82-e9ac-4fee-b1b5-3b3820a45868",
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(num_features):\n",
    "    start_col = i * time_steps\n",
    "    end_col = start_col + time_steps\n",
    "    X_rs[:, :, i] = df_X.iloc[:, start_col:end_col].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 305,
   "id": "e9068c90-7a47-4dd7-b06f-89cabc76c4ad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1432, 6, 5)"
      ]
     },
     "execution_count": 305,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_rs.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 306,
   "id": "5d3d8d5d-6458-42d4-ac73-221e51e75ea5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1145, 6, 5) (287, 6, 5) (1145,) (287,)\n"
     ]
    }
   ],
   "source": [
    "X_train_rsa, X_test_rs, y_train1a, y_test1 = train_test_split(X_rs, y, test_size = 0.2, random_state = 1)\n",
    "\n",
    "print(X_train_rsa.shape, X_test_rs.shape, y_train1a.shape, y_test1.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 307,
   "id": "ecccf797-26e2-4d1d-b5b0-472d744b9f0a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(916, 6, 5) (229, 6, 5) (916,) (229,)\n"
     ]
    }
   ],
   "source": [
    "X_train_rs, X_val_rs, y_train1, y_val1 = train_test_split(X_train_rsa, y_train1a, test_size = 0.2, random_state = 5)\n",
    "\n",
    "print(X_train_rs.shape, X_val_rs.shape, y_train1.shape, y_val1.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 308,
   "id": "507d1c3f-e031-4c52-a4c5-0eb427b7c98e",
   "metadata": {},
   "outputs": [],
   "source": [
    "#samples_test = X_test.shape[0]\n",
    "#X_test_rs = X_test[:samples_test*time_steps]\n",
    "#X_test_rs = np.reshape(X_test, (samples_test, time_steps, num_features))\n",
    "#X_test_rs.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "321ace18-9781-41bd-b250-a6da4bb38c6b",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_lstm2 = tf.keras.Sequential([\n",
    "    layers.LSTM(256, input_shape=(X_train_rs.shape[1], X_train_rs.shape[2]), return_sequences=True),\n",
    "    layers.LSTM(128, return_sequences=True),\n",
    "    layers.LSTM(64, return_sequences=True),\n",
    "    layers.LSTM(32),\n",
    "    layers.Dense(1, activation='linear')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 362,
   "id": "e25a9270-5980-488b-9aa3-c994451b66f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_lstm2.compile(loss='mse', optimizer=tf.keras.optimizers.Adam(learning_rate=1), metrics=[tfa.metrics.RSquare()])\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 363,
   "id": "7d6cbe53-dd6b-4b37-9292-8358a1a35570",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/400\n",
      "15/15 [==============================] - 15s 260ms/step - loss: 297437.4062 - r_square: -27.9273 - val_loss: 204674.0156 - val_r_square: -22.8370\n",
      "Epoch 2/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 119642.0625 - r_square: -10.6358 - val_loss: 70303.4297 - val_r_square: -7.1878\n",
      "Epoch 3/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 35886.5547 - r_square: -2.4902 - val_loss: 20477.9199 - val_r_square: -1.3849\n",
      "Epoch 4/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 12632.9150 - r_square: -0.2286 - val_loss: 9855.5713 - val_r_square: -0.1478\n",
      "Epoch 5/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10386.1318 - r_square: -0.0101 - val_loss: 8739.8574 - val_r_square: -0.0179\n",
      "Epoch 6/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10496.2041 - r_square: -0.0208 - val_loss: 8828.1875 - val_r_square: -0.0282\n",
      "Epoch 7/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10358.0498 - r_square: -0.0074 - val_loss: 9161.3955 - val_r_square: -0.0670\n",
      "Epoch 8/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10290.1328 - r_square: -7.7832e-04 - val_loss: 9381.6865 - val_r_square: -0.0926\n",
      "Epoch 9/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10284.9854 - r_square: -2.7430e-04 - val_loss: 9394.7139 - val_r_square: -0.0941\n",
      "Epoch 10/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10286.3506 - r_square: -4.0698e-04 - val_loss: 9360.3252 - val_r_square: -0.0901\n",
      "Epoch 11/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10285.9609 - r_square: -3.6561e-04 - val_loss: 9343.3857 - val_r_square: -0.0882\n",
      "Epoch 12/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10286.5156 - r_square: -4.2295e-04 - val_loss: 9380.1934 - val_r_square: -0.0924\n",
      "Epoch 13/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10284.7236 - r_square: -2.3854e-04 - val_loss: 9301.6416 - val_r_square: -0.0833\n",
      "Epoch 14/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10291.6035 - r_square: -9.1112e-04 - val_loss: 9236.1025 - val_r_square: -0.0757\n",
      "Epoch 15/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10291.4170 - r_square: -8.9967e-04 - val_loss: 9369.0117 - val_r_square: -0.0911\n",
      "Epoch 16/400\n",
      "15/15 [==============================] - 1s 86ms/step - loss: 10286.9531 - r_square: -4.6217e-04 - val_loss: 9311.4785 - val_r_square: -0.0844\n",
      "Epoch 17/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10294.8047 - r_square: -0.0012 - val_loss: 9272.0488 - val_r_square: -0.0799\n",
      "Epoch 18/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10294.3252 - r_square: -0.0012 - val_loss: 9393.8535 - val_r_square: -0.0940\n",
      "Epoch 19/400\n",
      "15/15 [==============================] - 2s 102ms/step - loss: 10298.0410 - r_square: -0.0015 - val_loss: 9434.6963 - val_r_square: -0.0988\n",
      "Epoch 20/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10301.2646 - r_square: -0.0018 - val_loss: 9130.9990 - val_r_square: -0.0634\n",
      "Epoch 21/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10287.0010 - r_square: -4.6003e-04 - val_loss: 9361.0176 - val_r_square: -0.0902\n",
      "Epoch 22/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10289.5010 - r_square: -7.1335e-04 - val_loss: 9316.7803 - val_r_square: -0.0851\n",
      "Epoch 23/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10287.2510 - r_square: -4.9460e-04 - val_loss: 9393.1494 - val_r_square: -0.0940\n",
      "Epoch 24/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10289.7168 - r_square: -7.3445e-04 - val_loss: 9336.2354 - val_r_square: -0.0873\n",
      "Epoch 25/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10287.1104 - r_square: -4.7410e-04 - val_loss: 9362.4414 - val_r_square: -0.0904\n",
      "Epoch 26/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10298.9502 - r_square: -0.0016 - val_loss: 9256.2754 - val_r_square: -0.0780\n",
      "Epoch 27/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10295.5615 - r_square: -0.0013 - val_loss: 9230.1025 - val_r_square: -0.0750\n",
      "Epoch 28/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10285.4346 - r_square: -3.1793e-04 - val_loss: 9432.1924 - val_r_square: -0.0985\n",
      "Epoch 29/400\n",
      "15/15 [==============================] - 1s 89ms/step - loss: 10313.5527 - r_square: -0.0030 - val_loss: 9314.9111 - val_r_square: -0.0848\n",
      "Epoch 30/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10357.5723 - r_square: -0.0073 - val_loss: 9526.6592 - val_r_square: -0.1095\n",
      "Epoch 31/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10291.7334 - r_square: -9.2709e-04 - val_loss: 9142.8965 - val_r_square: -0.0648\n",
      "Epoch 32/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10294.9160 - r_square: -0.0012 - val_loss: 9243.4697 - val_r_square: -0.0765\n",
      "Epoch 33/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10292.3857 - r_square: -9.9051e-04 - val_loss: 9501.2656 - val_r_square: -0.1065\n",
      "Epoch 34/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10283.9092 - r_square: -1.7977e-04 - val_loss: 9301.0938 - val_r_square: -0.0832\n",
      "Epoch 35/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10290.7100 - r_square: -8.3435e-04 - val_loss: 9193.3086 - val_r_square: -0.0707\n",
      "Epoch 36/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10294.6904 - r_square: -0.0012 - val_loss: 9411.0439 - val_r_square: -0.0960\n",
      "Epoch 37/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10328.6768 - r_square: -0.0045 - val_loss: 9176.6572 - val_r_square: -0.0687\n",
      "Epoch 38/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10289.9785 - r_square: -7.5984e-04 - val_loss: 9541.0869 - val_r_square: -0.1112\n",
      "Epoch 39/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10290.8076 - r_square: -8.4043e-04 - val_loss: 9330.6133 - val_r_square: -0.0867\n",
      "Epoch 40/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10290.7510 - r_square: -8.3160e-04 - val_loss: 9156.5293 - val_r_square: -0.0664\n",
      "Epoch 41/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10284.7578 - r_square: -2.4867e-04 - val_loss: 9353.4199 - val_r_square: -0.0893\n",
      "Epoch 42/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10311.6826 - r_square: -0.0029 - val_loss: 9320.4854 - val_r_square: -0.0855\n",
      "Epoch 43/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10285.1934 - r_square: -2.9099e-04 - val_loss: 9290.4990 - val_r_square: -0.0820\n",
      "Epoch 44/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10286.9502 - r_square: -4.5168e-04 - val_loss: 9290.5176 - val_r_square: -0.0820\n",
      "Epoch 45/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10298.7471 - r_square: -0.0016 - val_loss: 9252.1240 - val_r_square: -0.0775\n",
      "Epoch 46/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10279.2109 - r_square: 2.9081e-04 - val_loss: 9407.5635 - val_r_square: -0.0956\n",
      "Epoch 47/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10294.5195 - r_square: -0.0012 - val_loss: 9416.2148 - val_r_square: -0.0966\n",
      "Epoch 48/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10306.8184 - r_square: -0.0024 - val_loss: 9335.1133 - val_r_square: -0.0872\n",
      "Epoch 49/400\n",
      "15/15 [==============================] - 1s 86ms/step - loss: 10288.2578 - r_square: -6.0606e-04 - val_loss: 9333.2070 - val_r_square: -0.0870\n",
      "Epoch 50/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10306.1094 - r_square: -0.0023 - val_loss: 9483.7871 - val_r_square: -0.1045\n",
      "Epoch 51/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10282.6230 - r_square: -3.4332e-05 - val_loss: 9195.0684 - val_r_square: -0.0709\n",
      "Epoch 52/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10299.1426 - r_square: -0.0016 - val_loss: 9286.0332 - val_r_square: -0.0815\n",
      "Epoch 53/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10306.1914 - r_square: -0.0023 - val_loss: 9186.3672 - val_r_square: -0.0699\n",
      "Epoch 54/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10300.8438 - r_square: -0.0018 - val_loss: 9541.1855 - val_r_square: -0.1112\n",
      "Epoch 55/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10287.0195 - r_square: -4.7207e-04 - val_loss: 9215.9150 - val_r_square: -0.0733\n",
      "Epoch 56/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10291.8457 - r_square: -9.4140e-04 - val_loss: 9338.1387 - val_r_square: -0.0875\n",
      "Epoch 57/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10286.0742 - r_square: -3.7670e-04 - val_loss: 9263.2969 - val_r_square: -0.0788\n",
      "Epoch 58/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10300.7686 - r_square: -0.0018 - val_loss: 9293.6982 - val_r_square: -0.0824\n",
      "Epoch 59/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10287.0859 - r_square: -4.7517e-04 - val_loss: 9344.7969 - val_r_square: -0.0883\n",
      "Epoch 60/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10288.2139 - r_square: -5.8484e-04 - val_loss: 9292.2354 - val_r_square: -0.0822\n",
      "Epoch 61/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10291.2861 - r_square: -8.8704e-04 - val_loss: 9214.3262 - val_r_square: -0.0731\n",
      "Epoch 62/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10305.6865 - r_square: -0.0023 - val_loss: 9339.7627 - val_r_square: -0.0877\n",
      "Epoch 63/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10299.8018 - r_square: -0.0017 - val_loss: 9316.3955 - val_r_square: -0.0850\n",
      "Epoch 64/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10293.6279 - r_square: -0.0011 - val_loss: 9136.2666 - val_r_square: -0.0640\n",
      "Epoch 65/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10296.6016 - r_square: -0.0014 - val_loss: 9349.3555 - val_r_square: -0.0889\n",
      "Epoch 66/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10295.2676 - r_square: -0.0013 - val_loss: 9221.4268 - val_r_square: -0.0740\n",
      "Epoch 67/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10286.3652 - r_square: -3.9828e-04 - val_loss: 9320.0898 - val_r_square: -0.0854\n",
      "Epoch 68/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10306.2627 - r_square: -0.0023 - val_loss: 9337.7881 - val_r_square: -0.0875\n",
      "Epoch 69/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10303.6094 - r_square: -0.0021 - val_loss: 9220.8730 - val_r_square: -0.0739\n",
      "Epoch 70/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10361.3887 - r_square: -0.0077 - val_loss: 9337.7412 - val_r_square: -0.0875\n",
      "Epoch 71/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10316.6025 - r_square: -0.0033 - val_loss: 9258.4033 - val_r_square: -0.0783\n",
      "Epoch 72/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10301.7744 - r_square: -0.0019 - val_loss: 9478.0996 - val_r_square: -0.1038\n",
      "Epoch 73/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10294.8594 - r_square: -0.0012 - val_loss: 9336.0391 - val_r_square: -0.0873\n",
      "Epoch 74/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10292.9414 - r_square: -0.0010 - val_loss: 9312.2021 - val_r_square: -0.0845\n",
      "Epoch 75/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10293.7793 - r_square: -0.0011 - val_loss: 9308.8057 - val_r_square: -0.0841\n",
      "Epoch 76/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10317.9727 - r_square: -0.0035 - val_loss: 9214.4053 - val_r_square: -0.0731\n",
      "Epoch 77/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10276.9668 - r_square: 5.0902e-04 - val_loss: 9678.0635 - val_r_square: -0.1271\n",
      "Epoch 78/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10301.3740 - r_square: -0.0019 - val_loss: 9236.2334 - val_r_square: -0.0757\n",
      "Epoch 79/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10308.7773 - r_square: -0.0026 - val_loss: 9216.6924 - val_r_square: -0.0734\n",
      "Epoch 80/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10330.7373 - r_square: -0.0047 - val_loss: 9461.7861 - val_r_square: -0.1019\n",
      "Epoch 81/400\n",
      "15/15 [==============================] - 1s 88ms/step - loss: 10338.7881 - r_square: -0.0055 - val_loss: 9131.1592 - val_r_square: -0.0634\n",
      "Epoch 82/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10280.6299 - r_square: 1.5616e-04 - val_loss: 9440.3369 - val_r_square: -0.0995\n",
      "Epoch 83/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10325.9629 - r_square: -0.0043 - val_loss: 9366.0869 - val_r_square: -0.0908\n",
      "Epoch 84/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10293.0098 - r_square: -0.0011 - val_loss: 9181.2207 - val_r_square: -0.0693\n",
      "Epoch 85/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10309.2490 - r_square: -0.0026 - val_loss: 9255.7578 - val_r_square: -0.0780\n",
      "Epoch 86/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10312.1260 - r_square: -0.0029 - val_loss: 9144.4414 - val_r_square: -0.0650\n",
      "Epoch 87/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10309.9160 - r_square: -0.0027 - val_loss: 9194.4004 - val_r_square: -0.0708\n",
      "Epoch 88/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10294.0566 - r_square: -0.0012 - val_loss: 9361.2148 - val_r_square: -0.0902\n",
      "Epoch 89/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10288.9131 - r_square: -6.4600e-04 - val_loss: 9253.9004 - val_r_square: -0.0777\n",
      "Epoch 90/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10308.6543 - r_square: -0.0026 - val_loss: 9459.1406 - val_r_square: -0.1016\n",
      "Epoch 91/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10329.8174 - r_square: -0.0046 - val_loss: 9275.3867 - val_r_square: -0.0802\n",
      "Epoch 92/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10318.3975 - r_square: -0.0035 - val_loss: 9528.1836 - val_r_square: -0.1097\n",
      "Epoch 93/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10364.2285 - r_square: -0.0080 - val_loss: 9389.3438 - val_r_square: -0.0935\n",
      "Epoch 94/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10351.1748 - r_square: -0.0067 - val_loss: 9303.9590 - val_r_square: -0.0836\n",
      "Epoch 95/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10358.0488 - r_square: -0.0074 - val_loss: 9179.5547 - val_r_square: -0.0691\n",
      "Epoch 96/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10288.8164 - r_square: -6.5017e-04 - val_loss: 9468.1787 - val_r_square: -0.1027\n",
      "Epoch 97/400\n",
      "15/15 [==============================] - 1s 90ms/step - loss: 10299.5400 - r_square: -0.0017 - val_loss: 9217.6924 - val_r_square: -0.0735\n",
      "Epoch 98/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10355.7090 - r_square: -0.0071 - val_loss: 9282.6787 - val_r_square: -0.0811\n",
      "Epoch 99/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10321.4541 - r_square: -0.0038 - val_loss: 9331.3779 - val_r_square: -0.0868\n",
      "Epoch 100/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10307.3770 - r_square: -0.0024 - val_loss: 9313.0713 - val_r_square: -0.0846\n",
      "Epoch 101/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10297.7334 - r_square: -0.0015 - val_loss: 9446.3037 - val_r_square: -0.1001\n",
      "Epoch 102/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10293.7285 - r_square: -0.0011 - val_loss: 9136.4111 - val_r_square: -0.0641\n",
      "Epoch 103/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10357.5293 - r_square: -0.0073 - val_loss: 9465.4551 - val_r_square: -0.1024\n",
      "Epoch 104/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10313.6270 - r_square: -0.0031 - val_loss: 9198.8779 - val_r_square: -0.0713\n",
      "Epoch 105/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10317.6143 - r_square: -0.0034 - val_loss: 9328.6084 - val_r_square: -0.0864\n",
      "Epoch 106/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10303.3574 - r_square: -0.0021 - val_loss: 9113.5303 - val_r_square: -0.0614\n",
      "Epoch 107/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10347.3779 - r_square: -0.0063 - val_loss: 9464.1787 - val_r_square: -0.1022\n",
      "Epoch 108/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10277.6885 - r_square: 4.3541e-04 - val_loss: 9071.9121 - val_r_square: -0.0565\n",
      "Epoch 109/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10313.4473 - r_square: -0.0030 - val_loss: 9507.2012 - val_r_square: -0.1072\n",
      "Epoch 110/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10302.3438 - r_square: -0.0020 - val_loss: 9445.3701 - val_r_square: -0.1000\n",
      "Epoch 111/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10309.7041 - r_square: -0.0027 - val_loss: 9275.7256 - val_r_square: -0.0803\n",
      "Epoch 112/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10327.5303 - r_square: -0.0044 - val_loss: 9448.1816 - val_r_square: -0.1004\n",
      "Epoch 113/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10333.9805 - r_square: -0.0050 - val_loss: 9286.9463 - val_r_square: -0.0816\n",
      "Epoch 114/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10324.4717 - r_square: -0.0041 - val_loss: 9285.7656 - val_r_square: -0.0814\n",
      "Epoch 115/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10289.6953 - r_square: -7.2205e-04 - val_loss: 9079.8027 - val_r_square: -0.0575\n",
      "Epoch 116/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10330.9775 - r_square: -0.0047 - val_loss: 9360.8857 - val_r_square: -0.0902\n",
      "Epoch 117/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10357.2158 - r_square: -0.0073 - val_loss: 9183.1367 - val_r_square: -0.0695\n",
      "Epoch 118/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10303.2969 - r_square: -0.0021 - val_loss: 9303.2715 - val_r_square: -0.0835\n",
      "Epoch 119/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10299.1230 - r_square: -0.0016 - val_loss: 9315.3525 - val_r_square: -0.0849\n",
      "Epoch 120/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10324.7520 - r_square: -0.0041 - val_loss: 9248.7461 - val_r_square: -0.0771\n",
      "Epoch 121/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10323.5703 - r_square: -0.0040 - val_loss: 9468.6758 - val_r_square: -0.1028\n",
      "Epoch 122/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10324.8242 - r_square: -0.0041 - val_loss: 9319.7090 - val_r_square: -0.0854\n",
      "Epoch 123/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10319.5039 - r_square: -0.0036 - val_loss: 9455.9297 - val_r_square: -0.1013\n",
      "Epoch 124/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10314.8691 - r_square: -0.0032 - val_loss: 9144.3408 - val_r_square: -0.0650\n",
      "Epoch 125/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10340.2852 - r_square: -0.0056 - val_loss: 9253.5918 - val_r_square: -0.0777\n",
      "Epoch 126/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10317.6973 - r_square: -0.0034 - val_loss: 9239.7979 - val_r_square: -0.0761\n",
      "Epoch 127/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10304.8867 - r_square: -0.0022 - val_loss: 9222.7246 - val_r_square: -0.0741\n",
      "Epoch 128/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10296.2002 - r_square: -0.0014 - val_loss: 9375.5889 - val_r_square: -0.0919\n",
      "Epoch 129/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10307.4502 - r_square: -0.0024 - val_loss: 9283.7354 - val_r_square: -0.0812\n",
      "Epoch 130/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10355.5918 - r_square: -0.0071 - val_loss: 9257.6152 - val_r_square: -0.0782\n",
      "Epoch 131/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10372.1328 - r_square: -0.0087 - val_loss: 9451.6699 - val_r_square: -0.1008\n",
      "Epoch 132/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10487.1602 - r_square: -0.0199 - val_loss: 9178.5381 - val_r_square: -0.0690\n",
      "Epoch 133/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10290.2402 - r_square: -7.8189e-04 - val_loss: 9464.0996 - val_r_square: -0.1022\n",
      "Epoch 134/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10312.4668 - r_square: -0.0029 - val_loss: 9304.0400 - val_r_square: -0.0836\n",
      "Epoch 135/400\n",
      "15/15 [==============================] - 1s 91ms/step - loss: 10316.8418 - r_square: -0.0034 - val_loss: 9266.8740 - val_r_square: -0.0792\n",
      "Epoch 136/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10324.1699 - r_square: -0.0041 - val_loss: 9391.6455 - val_r_square: -0.0938\n",
      "Epoch 137/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10298.9482 - r_square: -0.0016 - val_loss: 9086.4326 - val_r_square: -0.0582\n",
      "Epoch 138/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10300.8340 - r_square: -0.0018 - val_loss: 9452.2568 - val_r_square: -0.1008\n",
      "Epoch 139/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10303.5254 - r_square: -0.0021 - val_loss: 9459.0361 - val_r_square: -0.1016\n",
      "Epoch 140/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10291.4053 - r_square: -8.9860e-04 - val_loss: 9283.4385 - val_r_square: -0.0812\n",
      "Epoch 141/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10305.2705 - r_square: -0.0022 - val_loss: 9191.5771 - val_r_square: -0.0705\n",
      "Epoch 142/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10348.2012 - r_square: -0.0064 - val_loss: 9519.9316 - val_r_square: -0.1087\n",
      "Epoch 143/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10420.2900 - r_square: -0.0134 - val_loss: 9061.8359 - val_r_square: -0.0554\n",
      "Epoch 144/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10356.1104 - r_square: -0.0072 - val_loss: 9595.6309 - val_r_square: -0.1175\n",
      "Epoch 145/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10308.0859 - r_square: -0.0025 - val_loss: 9181.5459 - val_r_square: -0.0693\n",
      "Epoch 146/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10340.9814 - r_square: -0.0057 - val_loss: 9475.6074 - val_r_square: -0.1036\n",
      "Epoch 147/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10372.5811 - r_square: -0.0088 - val_loss: 9075.7852 - val_r_square: -0.0570\n",
      "Epoch 148/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10342.9795 - r_square: -0.0059 - val_loss: 9505.4717 - val_r_square: -0.1070\n",
      "Epoch 149/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10297.3672 - r_square: -0.0015 - val_loss: 9328.9268 - val_r_square: -0.0865\n",
      "Epoch 150/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10364.1729 - r_square: -0.0080 - val_loss: 9757.9697 - val_r_square: -0.1364\n",
      "Epoch 151/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10332.2490 - r_square: -0.0049 - val_loss: 9300.3203 - val_r_square: -0.0831\n",
      "Epoch 152/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10300.6719 - r_square: -0.0018 - val_loss: 9302.8076 - val_r_square: -0.0834\n",
      "Epoch 153/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10297.6113 - r_square: -0.0015 - val_loss: 9229.2275 - val_r_square: -0.0749\n",
      "Epoch 154/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10313.1973 - r_square: -0.0030 - val_loss: 9137.0986 - val_r_square: -0.0641\n",
      "Epoch 155/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10347.5586 - r_square: -0.0064 - val_loss: 9587.2812 - val_r_square: -0.1166\n",
      "Epoch 156/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10303.1133 - r_square: -0.0020 - val_loss: 9084.6865 - val_r_square: -0.0580\n",
      "Epoch 157/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10320.8672 - r_square: -0.0038 - val_loss: 9428.3516 - val_r_square: -0.0981\n",
      "Epoch 158/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10324.7578 - r_square: -0.0041 - val_loss: 9056.4727 - val_r_square: -0.0547\n",
      "Epoch 159/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10307.4111 - r_square: -0.0024 - val_loss: 9478.5146 - val_r_square: -0.1039\n",
      "Epoch 160/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10297.0498 - r_square: -0.0014 - val_loss: 9261.3047 - val_r_square: -0.0786\n",
      "Epoch 161/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10313.6416 - r_square: -0.0031 - val_loss: 9557.7207 - val_r_square: -0.1131\n",
      "Epoch 162/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10330.6475 - r_square: -0.0047 - val_loss: 9026.0264 - val_r_square: -0.0512\n",
      "Epoch 163/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10289.0732 - r_square: -6.6841e-04 - val_loss: 9616.5273 - val_r_square: -0.1200\n",
      "Epoch 164/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10325.0918 - r_square: -0.0042 - val_loss: 9194.3682 - val_r_square: -0.0708\n",
      "Epoch 165/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10302.9072 - r_square: -0.0020 - val_loss: 9163.1182 - val_r_square: -0.0672\n",
      "Epoch 166/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10289.4102 - r_square: -7.0119e-04 - val_loss: 9703.2393 - val_r_square: -0.1301\n",
      "Epoch 167/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10365.8164 - r_square: -0.0081 - val_loss: 9177.3066 - val_r_square: -0.0688\n",
      "Epoch 168/400\n",
      "15/15 [==============================] - 1s 95ms/step - loss: 10302.1689 - r_square: -0.0019 - val_loss: 9335.9102 - val_r_square: -0.0873\n",
      "Epoch 169/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10333.1738 - r_square: -0.0050 - val_loss: 9199.9717 - val_r_square: -0.0715\n",
      "Epoch 170/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10293.5010 - r_square: -0.0011 - val_loss: 9541.7061 - val_r_square: -0.1113\n",
      "Epoch 171/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10294.3652 - r_square: -0.0012 - val_loss: 9045.4131 - val_r_square: -0.0535\n",
      "Epoch 172/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10319.2529 - r_square: -0.0036 - val_loss: 9580.3574 - val_r_square: -0.1158\n",
      "Epoch 173/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10316.6123 - r_square: -0.0033 - val_loss: 9249.0918 - val_r_square: -0.0772\n",
      "Epoch 174/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10415.4912 - r_square: -0.0130 - val_loss: 8922.3945 - val_r_square: -0.0391\n",
      "Epoch 175/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10332.9326 - r_square: -0.0049 - val_loss: 9786.2031 - val_r_square: -0.1397\n",
      "Epoch 176/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10332.6553 - r_square: -0.0049 - val_loss: 9202.3779 - val_r_square: -0.0717\n",
      "Epoch 177/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10314.5293 - r_square: -0.0031 - val_loss: 9713.0527 - val_r_square: -0.1312\n",
      "Epoch 178/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10377.2842 - r_square: -0.0092 - val_loss: 8752.3906 - val_r_square: -0.0193\n",
      "Epoch 179/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10535.7402 - r_square: -0.0247 - val_loss: 10059.7754 - val_r_square: -0.1716\n",
      "Epoch 180/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10401.5840 - r_square: -0.0116 - val_loss: 9194.5928 - val_r_square: -0.0708\n",
      "Epoch 181/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10320.2910 - r_square: -0.0037 - val_loss: 9372.3721 - val_r_square: -0.0915\n",
      "Epoch 182/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10331.5957 - r_square: -0.0048 - val_loss: 9730.6973 - val_r_square: -0.1333\n",
      "Epoch 183/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10335.2803 - r_square: -0.0052 - val_loss: 9160.7432 - val_r_square: -0.0669\n",
      "Epoch 184/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10353.5527 - r_square: -0.0069 - val_loss: 9061.5303 - val_r_square: -0.0553\n",
      "Epoch 185/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10307.4258 - r_square: -0.0024 - val_loss: 9817.3076 - val_r_square: -0.1434\n",
      "Epoch 186/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10341.9248 - r_square: -0.0058 - val_loss: 8916.2568 - val_r_square: -0.0384\n",
      "Epoch 187/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10441.1973 - r_square: -0.0155 - val_loss: 9461.2480 - val_r_square: -0.1019\n",
      "Epoch 188/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10308.1602 - r_square: -0.0025 - val_loss: 9250.5967 - val_r_square: -0.0774\n",
      "Epoch 189/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10294.2607 - r_square: -0.0012 - val_loss: 9698.1348 - val_r_square: -0.1295\n",
      "Epoch 190/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10329.3213 - r_square: -0.0046 - val_loss: 8936.1963 - val_r_square: -0.0407\n",
      "Epoch 191/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10424.9258 - r_square: -0.0139 - val_loss: 9816.3535 - val_r_square: -0.1432\n",
      "Epoch 192/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10335.8203 - r_square: -0.0052 - val_loss: 9414.8438 - val_r_square: -0.0965\n",
      "Epoch 193/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10332.0029 - r_square: -0.0048 - val_loss: 9359.3906 - val_r_square: -0.0900\n",
      "Epoch 194/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10306.1035 - r_square: -0.0023 - val_loss: 9425.3799 - val_r_square: -0.0977\n",
      "Epoch 195/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10307.1592 - r_square: -0.0024 - val_loss: 9168.5410 - val_r_square: -0.0678\n",
      "Epoch 196/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10298.7383 - r_square: -0.0016 - val_loss: 9259.8408 - val_r_square: -0.0784\n",
      "Epoch 197/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10327.5186 - r_square: -0.0044 - val_loss: 9885.0195 - val_r_square: -0.1512\n",
      "Epoch 198/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10312.3066 - r_square: -0.0029 - val_loss: 8902.4746 - val_r_square: -0.0368\n",
      "Epoch 199/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10390.6406 - r_square: -0.0105 - val_loss: 10295.7422 - val_r_square: -0.1991\n",
      "Epoch 200/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10418.6279 - r_square: -0.0133 - val_loss: 8805.1152 - val_r_square: -0.0255\n",
      "Epoch 201/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10422.5215 - r_square: -0.0136 - val_loss: 9714.0020 - val_r_square: -0.1313\n",
      "Epoch 202/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10281.1934 - r_square: 9.4593e-05 - val_loss: 9050.6348 - val_r_square: -0.0541\n",
      "Epoch 203/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10307.6973 - r_square: -0.0025 - val_loss: 9718.0801 - val_r_square: -0.1318\n",
      "Epoch 204/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10330.7656 - r_square: -0.0047 - val_loss: 8992.5391 - val_r_square: -0.0473\n",
      "Epoch 205/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10351.7998 - r_square: -0.0068 - val_loss: 9170.3125 - val_r_square: -0.0680\n",
      "Epoch 206/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10331.6924 - r_square: -0.0048 - val_loss: 9281.6650 - val_r_square: -0.0810\n",
      "Epoch 207/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10293.2744 - r_square: -0.0011 - val_loss: 9266.2285 - val_r_square: -0.0792\n",
      "Epoch 208/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10358.6318 - r_square: -0.0074 - val_loss: 9402.0869 - val_r_square: -0.0950\n",
      "Epoch 209/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10329.7471 - r_square: -0.0046 - val_loss: 9446.0566 - val_r_square: -0.1001\n",
      "Epoch 210/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10305.5068 - r_square: -0.0023 - val_loss: 9121.7422 - val_r_square: -0.0623\n",
      "Epoch 211/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10313.8389 - r_square: -0.0031 - val_loss: 9326.7334 - val_r_square: -0.0862\n",
      "Epoch 212/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10297.4775 - r_square: -0.0015 - val_loss: 9266.9326 - val_r_square: -0.0793\n",
      "Epoch 213/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10308.1172 - r_square: -0.0025 - val_loss: 9066.6416 - val_r_square: -0.0559\n",
      "Epoch 214/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10331.4150 - r_square: -0.0048 - val_loss: 9301.2334 - val_r_square: -0.0833\n",
      "Epoch 215/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10363.1738 - r_square: -0.0079 - val_loss: 9773.4336 - val_r_square: -0.1382\n",
      "Epoch 216/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10391.9727 - r_square: -0.0107 - val_loss: 9744.4922 - val_r_square: -0.1349\n",
      "Epoch 217/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10326.9834 - r_square: -0.0044 - val_loss: 9075.5820 - val_r_square: -0.0570\n",
      "Epoch 218/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10319.3613 - r_square: -0.0036 - val_loss: 9412.9170 - val_r_square: -0.0963\n",
      "Epoch 219/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10316.3945 - r_square: -0.0033 - val_loss: 9061.5674 - val_r_square: -0.0553\n",
      "Epoch 220/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10307.7559 - r_square: -0.0025 - val_loss: 9509.4033 - val_r_square: -0.1075\n",
      "Epoch 221/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10312.6934 - r_square: -0.0030 - val_loss: 9283.1230 - val_r_square: -0.0811\n",
      "Epoch 222/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10351.9619 - r_square: -0.0068 - val_loss: 9599.5283 - val_r_square: -0.1180\n",
      "Epoch 223/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10358.7705 - r_square: -0.0075 - val_loss: 9207.9619 - val_r_square: -0.0724\n",
      "Epoch 224/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10379.6270 - r_square: -0.0095 - val_loss: 9152.9004 - val_r_square: -0.0660\n",
      "Epoch 225/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10303.9834 - r_square: -0.0021 - val_loss: 9256.1777 - val_r_square: -0.0780\n",
      "Epoch 226/400\n",
      "15/15 [==============================] - 1s 89ms/step - loss: 10339.5996 - r_square: -0.0056 - val_loss: 9331.6797 - val_r_square: -0.0868\n",
      "Epoch 227/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10311.7559 - r_square: -0.0029 - val_loss: 9108.8770 - val_r_square: -0.0608\n",
      "Epoch 228/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10467.9902 - r_square: -0.0181 - val_loss: 10157.0195 - val_r_square: -0.1829\n",
      "Epoch 229/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10396.8818 - r_square: -0.0111 - val_loss: 9488.7559 - val_r_square: -0.1051\n",
      "Epoch 230/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10334.1982 - r_square: -0.0051 - val_loss: 8943.4229 - val_r_square: -0.0416\n",
      "Epoch 231/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10378.4561 - r_square: -0.0094 - val_loss: 9710.5088 - val_r_square: -0.1309\n",
      "Epoch 232/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10306.1777 - r_square: -0.0023 - val_loss: 9186.1816 - val_r_square: -0.0699\n",
      "Epoch 233/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10304.9619 - r_square: -0.0022 - val_loss: 9362.2754 - val_r_square: -0.0904\n",
      "Epoch 234/400\n",
      "15/15 [==============================] - 1s 90ms/step - loss: 10305.4512 - r_square: -0.0023 - val_loss: 9746.2100 - val_r_square: -0.1351\n",
      "Epoch 235/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10400.1309 - r_square: -0.0115 - val_loss: 9073.2168 - val_r_square: -0.0567\n",
      "Epoch 236/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10405.0479 - r_square: -0.0119 - val_loss: 9253.2422 - val_r_square: -0.0777\n",
      "Epoch 237/400\n",
      "15/15 [==============================] - 1s 78ms/step - loss: 10446.4541 - r_square: -0.0160 - val_loss: 10784.7422 - val_r_square: -0.2560\n",
      "Epoch 238/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10594.4053 - r_square: -0.0304 - val_loss: 8936.3633 - val_r_square: -0.0408\n",
      "Epoch 239/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10421.0146 - r_square: -0.0135 - val_loss: 9046.3916 - val_r_square: -0.0536\n",
      "Epoch 240/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10332.9512 - r_square: -0.0049 - val_loss: 9335.8145 - val_r_square: -0.0873\n",
      "Epoch 241/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10330.8145 - r_square: -0.0047 - val_loss: 9245.2383 - val_r_square: -0.0767\n",
      "Epoch 242/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10462.9111 - r_square: -0.0176 - val_loss: 9846.4854 - val_r_square: -0.1468\n",
      "Epoch 243/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10444.4941 - r_square: -0.0158 - val_loss: 9692.2578 - val_r_square: -0.1288\n",
      "Epoch 244/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10433.1377 - r_square: -0.0147 - val_loss: 9100.2432 - val_r_square: -0.0598\n",
      "Epoch 245/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10327.3340 - r_square: -0.0044 - val_loss: 9871.2754 - val_r_square: -0.1496\n",
      "Epoch 246/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10308.2012 - r_square: -0.0025 - val_loss: 9192.2031 - val_r_square: -0.0706\n",
      "Epoch 247/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10312.9219 - r_square: -0.0030 - val_loss: 9195.1201 - val_r_square: -0.0709\n",
      "Epoch 248/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10300.8037 - r_square: -0.0018 - val_loss: 9468.9414 - val_r_square: -0.1028\n",
      "Epoch 249/400\n",
      "15/15 [==============================] - 1s 88ms/step - loss: 10318.4131 - r_square: -0.0035 - val_loss: 9103.8975 - val_r_square: -0.0603\n",
      "Epoch 250/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10328.7842 - r_square: -0.0045 - val_loss: 9231.2979 - val_r_square: -0.0751\n",
      "Epoch 251/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10281.4736 - r_square: 7.0691e-05 - val_loss: 9598.4219 - val_r_square: -0.1179\n",
      "Epoch 252/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10350.8135 - r_square: -0.0067 - val_loss: 9389.8975 - val_r_square: -0.0936\n",
      "Epoch 253/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10312.4717 - r_square: -0.0029 - val_loss: 9674.0938 - val_r_square: -0.1267\n",
      "Epoch 254/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10337.4629 - r_square: -0.0054 - val_loss: 8900.7734 - val_r_square: -0.0366\n",
      "Epoch 255/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10343.6797 - r_square: -0.0060 - val_loss: 9263.8789 - val_r_square: -0.0789\n",
      "Epoch 256/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10315.5410 - r_square: -0.0032 - val_loss: 9217.1797 - val_r_square: -0.0735\n",
      "Epoch 257/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10328.7480 - r_square: -0.0045 - val_loss: 8854.4844 - val_r_square: -0.0312\n",
      "Epoch 258/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10331.6406 - r_square: -0.0048 - val_loss: 10390.8174 - val_r_square: -0.2101\n",
      "Epoch 259/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10453.8115 - r_square: -0.0167 - val_loss: 9039.6865 - val_r_square: -0.0528\n",
      "Epoch 260/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10318.6865 - r_square: -0.0036 - val_loss: 9185.2012 - val_r_square: -0.0697\n",
      "Epoch 261/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10335.6660 - r_square: -0.0052 - val_loss: 9675.9824 - val_r_square: -0.1269\n",
      "Epoch 262/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10355.6533 - r_square: -0.0071 - val_loss: 9254.7998 - val_r_square: -0.0778\n",
      "Epoch 263/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10378.6885 - r_square: -0.0094 - val_loss: 9063.3848 - val_r_square: -0.0555\n",
      "Epoch 264/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10354.9043 - r_square: -0.0071 - val_loss: 9018.4170 - val_r_square: -0.0503\n",
      "Epoch 265/400\n",
      "15/15 [==============================] - 1s 100ms/step - loss: 10346.6504 - r_square: -0.0063 - val_loss: 9000.0498 - val_r_square: -0.0482\n",
      "Epoch 266/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10404.0088 - r_square: -0.0118 - val_loss: 10759.0068 - val_r_square: -0.2530\n",
      "Epoch 267/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10533.1729 - r_square: -0.0244 - val_loss: 9375.7744 - val_r_square: -0.0919\n",
      "Epoch 268/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10324.6426 - r_square: -0.0041 - val_loss: 9570.1816 - val_r_square: -0.1146\n",
      "Epoch 269/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10290.8418 - r_square: -8.3351e-04 - val_loss: 9231.6904 - val_r_square: -0.0752\n",
      "Epoch 270/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10297.6455 - r_square: -0.0015 - val_loss: 9408.0703 - val_r_square: -0.0957\n",
      "Epoch 271/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10303.2695 - r_square: -0.0020 - val_loss: 9363.0518 - val_r_square: -0.0905\n",
      "Epoch 272/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10328.9268 - r_square: -0.0046 - val_loss: 8926.1230 - val_r_square: -0.0396\n",
      "Epoch 273/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10437.7793 - r_square: -0.0151 - val_loss: 9493.7520 - val_r_square: -0.1057\n",
      "Epoch 274/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10370.0127 - r_square: -0.0085 - val_loss: 10229.9814 - val_r_square: -0.1914\n",
      "Epoch 275/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10506.5850 - r_square: -0.0218 - val_loss: 9598.9346 - val_r_square: -0.1179\n",
      "Epoch 276/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10338.0449 - r_square: -0.0054 - val_loss: 9042.1191 - val_r_square: -0.0531\n",
      "Epoch 277/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10311.9844 - r_square: -0.0029 - val_loss: 9240.3604 - val_r_square: -0.0762\n",
      "Epoch 278/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10374.1445 - r_square: -0.0089 - val_loss: 9716.0703 - val_r_square: -0.1316\n",
      "Epoch 279/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10363.0254 - r_square: -0.0079 - val_loss: 9147.7793 - val_r_square: -0.0654\n",
      "Epoch 280/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10332.0527 - r_square: -0.0048 - val_loss: 8825.4551 - val_r_square: -0.0278\n",
      "Epoch 281/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10501.5146 - r_square: -0.0213 - val_loss: 9847.9521 - val_r_square: -0.1469\n",
      "Epoch 282/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10354.4404 - r_square: -0.0070 - val_loss: 9603.5732 - val_r_square: -0.1185\n",
      "Epoch 283/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10323.0400 - r_square: -0.0040 - val_loss: 9069.8926 - val_r_square: -0.0563\n",
      "Epoch 284/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10317.6377 - r_square: -0.0034 - val_loss: 9961.0166 - val_r_square: -0.1601\n",
      "Epoch 285/400\n",
      "15/15 [==============================] - 1s 94ms/step - loss: 10398.4375 - r_square: -0.0113 - val_loss: 9139.4590 - val_r_square: -0.0644\n",
      "Epoch 286/400\n",
      "15/15 [==============================] - 1s 100ms/step - loss: 10325.2090 - r_square: -0.0042 - val_loss: 8972.6904 - val_r_square: -0.0450\n",
      "Epoch 287/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10283.0635 - r_square: -8.0466e-05 - val_loss: 10076.4873 - val_r_square: -0.1735\n",
      "Epoch 288/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10370.3906 - r_square: -0.0086 - val_loss: 9286.2881 - val_r_square: -0.0815\n",
      "Epoch 289/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10289.2227 - r_square: -6.7949e-04 - val_loss: 9504.6396 - val_r_square: -0.1069\n",
      "Epoch 290/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10298.7939 - r_square: -0.0016 - val_loss: 9327.8594 - val_r_square: -0.0864\n",
      "Epoch 291/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10310.5566 - r_square: -0.0028 - val_loss: 9224.0537 - val_r_square: -0.0743\n",
      "Epoch 292/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10317.0312 - r_square: -0.0034 - val_loss: 9494.8340 - val_r_square: -0.1058\n",
      "Epoch 293/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10308.4600 - r_square: -0.0026 - val_loss: 9892.3799 - val_r_square: -0.1521\n",
      "Epoch 294/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10488.6104 - r_square: -0.0201 - val_loss: 8664.4043 - val_r_square: -0.0091\n",
      "Epoch 295/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10373.1846 - r_square: -0.0088 - val_loss: 10880.4902 - val_r_square: -0.2672\n",
      "Epoch 296/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10404.2314 - r_square: -0.0119 - val_loss: 9047.3301 - val_r_square: -0.0537\n",
      "Epoch 297/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10304.8711 - r_square: -0.0022 - val_loss: 9322.9424 - val_r_square: -0.0858\n",
      "Epoch 298/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10337.0586 - r_square: -0.0053 - val_loss: 9392.9521 - val_r_square: -0.0939\n",
      "Epoch 299/400\n",
      "15/15 [==============================] - 1s 91ms/step - loss: 10311.7207 - r_square: -0.0029 - val_loss: 9376.7861 - val_r_square: -0.0920\n",
      "Epoch 300/400\n",
      "15/15 [==============================] - 1s 86ms/step - loss: 10308.1562 - r_square: -0.0025 - val_loss: 9109.7002 - val_r_square: -0.0609\n",
      "Epoch 301/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10305.3857 - r_square: -0.0023 - val_loss: 9375.3115 - val_r_square: -0.0919\n",
      "Epoch 302/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10296.5430 - r_square: -0.0014 - val_loss: 9305.1006 - val_r_square: -0.0837\n",
      "Epoch 303/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10294.7676 - r_square: -0.0012 - val_loss: 9733.6143 - val_r_square: -0.1336\n",
      "Epoch 304/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10453.4570 - r_square: -0.0166 - val_loss: 9005.8604 - val_r_square: -0.0489\n",
      "Epoch 305/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10500.0518 - r_square: -0.0212 - val_loss: 8841.9785 - val_r_square: -0.0298\n",
      "Epoch 306/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10311.6865 - r_square: -0.0029 - val_loss: 9304.0732 - val_r_square: -0.0836\n",
      "Epoch 307/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10376.0684 - r_square: -0.0091 - val_loss: 10690.6943 - val_r_square: -0.2451\n",
      "Epoch 308/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10620.2393 - r_square: -0.0329 - val_loss: 9029.0273 - val_r_square: -0.0515\n",
      "Epoch 309/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10335.8047 - r_square: -0.0052 - val_loss: 9108.5039 - val_r_square: -0.0608\n",
      "Epoch 310/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10360.6074 - r_square: -0.0076 - val_loss: 9917.7354 - val_r_square: -0.1551\n",
      "Epoch 311/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10359.3633 - r_square: -0.0075 - val_loss: 9508.9326 - val_r_square: -0.1074\n",
      "Epoch 312/400\n",
      "15/15 [==============================] - 1s 94ms/step - loss: 10391.7021 - r_square: -0.0106 - val_loss: 8993.9814 - val_r_square: -0.0475\n",
      "Epoch 313/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10433.0615 - r_square: -0.0147 - val_loss: 8749.2373 - val_r_square: -0.0190\n",
      "Epoch 314/400\n",
      "15/15 [==============================] - 1s 79ms/step - loss: 10351.1855 - r_square: -0.0067 - val_loss: 9367.7832 - val_r_square: -0.0910\n",
      "Epoch 315/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10310.5674 - r_square: -0.0027 - val_loss: 9705.9736 - val_r_square: -0.1304\n",
      "Epoch 316/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10325.1016 - r_square: -0.0042 - val_loss: 9045.1514 - val_r_square: -0.0534\n",
      "Epoch 317/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10315.6787 - r_square: -0.0033 - val_loss: 9887.5410 - val_r_square: -0.1515\n",
      "Epoch 318/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10395.0547 - r_square: -0.0110 - val_loss: 9690.8330 - val_r_square: -0.1286\n",
      "Epoch 319/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10293.7578 - r_square: -0.0011 - val_loss: 9143.7344 - val_r_square: -0.0649\n",
      "Epoch 320/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10299.9609 - r_square: -0.0017 - val_loss: 9101.0811 - val_r_square: -0.0599\n",
      "Epoch 321/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10323.2715 - r_square: -0.0040 - val_loss: 9990.1055 - val_r_square: -0.1635\n",
      "Epoch 322/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10561.7949 - r_square: -0.0272 - val_loss: 9245.8496 - val_r_square: -0.0768\n",
      "Epoch 323/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10309.4326 - r_square: -0.0026 - val_loss: 9949.4551 - val_r_square: -0.1587\n",
      "Epoch 324/400\n",
      "15/15 [==============================] - 1s 92ms/step - loss: 10470.4160 - r_square: -0.0183 - val_loss: 8757.4102 - val_r_square: -0.0199\n",
      "Epoch 325/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10439.0439 - r_square: -0.0153 - val_loss: 9478.3115 - val_r_square: -0.1039\n",
      "Epoch 326/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10478.4824 - r_square: -0.0191 - val_loss: 9892.0020 - val_r_square: -0.1521\n",
      "Epoch 327/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10513.6270 - r_square: -0.0225 - val_loss: 10158.8682 - val_r_square: -0.1831\n",
      "Epoch 328/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10388.2754 - r_square: -0.0103 - val_loss: 9559.3389 - val_r_square: -0.1133\n",
      "Epoch 329/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10433.6133 - r_square: -0.0147 - val_loss: 9392.0020 - val_r_square: -0.0938\n",
      "Epoch 330/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10370.6094 - r_square: -0.0086 - val_loss: 9463.1445 - val_r_square: -0.1021\n",
      "Epoch 331/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10355.7793 - r_square: -0.0072 - val_loss: 9427.4453 - val_r_square: -0.0979\n",
      "Epoch 332/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10337.3076 - r_square: -0.0054 - val_loss: 9442.7988 - val_r_square: -0.0997\n",
      "Epoch 333/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10320.2646 - r_square: -0.0037 - val_loss: 8872.2832 - val_r_square: -0.0333\n",
      "Epoch 334/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10393.1123 - r_square: -0.0108 - val_loss: 8989.3545 - val_r_square: -0.0469\n",
      "Epoch 335/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10331.8301 - r_square: -0.0048 - val_loss: 10138.3057 - val_r_square: -0.1807\n",
      "Epoch 336/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10371.7148 - r_square: -0.0087 - val_loss: 9229.1553 - val_r_square: -0.0749\n",
      "Epoch 337/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10331.2598 - r_square: -0.0048 - val_loss: 8911.2559 - val_r_square: -0.0378\n",
      "Epoch 338/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10355.4717 - r_square: -0.0071 - val_loss: 10155.5488 - val_r_square: -0.1827\n",
      "Epoch 339/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10447.7510 - r_square: -0.0161 - val_loss: 9480.3086 - val_r_square: -0.1041\n",
      "Epoch 340/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10311.3633 - r_square: -0.0028 - val_loss: 9396.1904 - val_r_square: -0.0943\n",
      "Epoch 341/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10328.9219 - r_square: -0.0045 - val_loss: 9240.8975 - val_r_square: -0.0762\n",
      "Epoch 342/400\n",
      "15/15 [==============================] - 1s 96ms/step - loss: 10385.2334 - r_square: -0.0100 - val_loss: 9605.3535 - val_r_square: -0.1187\n",
      "Epoch 343/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10314.7793 - r_square: -0.0032 - val_loss: 9422.5674 - val_r_square: -0.0974\n",
      "Epoch 344/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10475.1582 - r_square: -0.0188 - val_loss: 9299.8545 - val_r_square: -0.0831\n",
      "Epoch 345/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10292.4229 - r_square: -0.0010 - val_loss: 9322.2900 - val_r_square: -0.0857\n",
      "Epoch 346/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10319.5293 - r_square: -0.0036 - val_loss: 9624.0527 - val_r_square: -0.1208\n",
      "Epoch 347/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10354.0303 - r_square: -0.0070 - val_loss: 9975.7285 - val_r_square: -0.1618\n",
      "Epoch 348/400\n",
      "15/15 [==============================] - 1s 95ms/step - loss: 10410.0439 - r_square: -0.0124 - val_loss: 8782.7529 - val_r_square: -0.0229\n",
      "Epoch 349/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10459.2979 - r_square: -0.0172 - val_loss: 8976.6182 - val_r_square: -0.0454\n",
      "Epoch 350/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10436.6260 - r_square: -0.0150 - val_loss: 10411.5146 - val_r_square: -0.2126\n",
      "Epoch 351/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10456.2207 - r_square: -0.0169 - val_loss: 9129.5908 - val_r_square: -0.0633\n",
      "Epoch 352/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10339.1680 - r_square: -0.0055 - val_loss: 9016.3740 - val_r_square: -0.0501\n",
      "Epoch 353/400\n",
      "15/15 [==============================] - 1s 94ms/step - loss: 10379.6455 - r_square: -0.0095 - val_loss: 9187.2725 - val_r_square: -0.0700\n",
      "Epoch 354/400\n",
      "15/15 [==============================] - 1s 85ms/step - loss: 10313.1846 - r_square: -0.0030 - val_loss: 9496.9062 - val_r_square: -0.1060\n",
      "Epoch 355/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10329.1055 - r_square: -0.0046 - val_loss: 10047.1787 - val_r_square: -0.1701\n",
      "Epoch 356/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10366.1982 - r_square: -0.0082 - val_loss: 9014.8545 - val_r_square: -0.0499\n",
      "Epoch 357/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10356.2119 - r_square: -0.0072 - val_loss: 9111.5557 - val_r_square: -0.0612\n",
      "Epoch 358/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10309.3867 - r_square: -0.0026 - val_loss: 9098.8154 - val_r_square: -0.0597\n",
      "Epoch 359/400\n",
      "15/15 [==============================] - 1s 94ms/step - loss: 10311.8760 - r_square: -0.0029 - val_loss: 9663.8086 - val_r_square: -0.1255\n",
      "Epoch 360/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10299.9463 - r_square: -0.0017 - val_loss: 8821.4697 - val_r_square: -0.0274\n",
      "Epoch 361/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10333.2705 - r_square: -0.0050 - val_loss: 10017.9062 - val_r_square: -0.1667\n",
      "Epoch 362/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10287.6758 - r_square: -5.3251e-04 - val_loss: 8924.6768 - val_r_square: -0.0394\n",
      "Epoch 363/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10296.7344 - r_square: -0.0014 - val_loss: 9524.7930 - val_r_square: -0.1093\n",
      "Epoch 364/400\n",
      "15/15 [==============================] - 1s 94ms/step - loss: 10315.7891 - r_square: -0.0033 - val_loss: 9814.2148 - val_r_square: -0.1430\n",
      "Epoch 365/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10431.1562 - r_square: -0.0145 - val_loss: 9126.9277 - val_r_square: -0.0630\n",
      "Epoch 366/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10350.7393 - r_square: -0.0067 - val_loss: 9226.1602 - val_r_square: -0.0745\n",
      "Epoch 367/400\n",
      "15/15 [==============================] - 1s 80ms/step - loss: 10306.2666 - r_square: -0.0023 - val_loss: 9638.1641 - val_r_square: -0.1225\n",
      "Epoch 368/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10363.5107 - r_square: -0.0079 - val_loss: 9109.7031 - val_r_square: -0.0609\n",
      "Epoch 369/400\n",
      "15/15 [==============================] - 1s 93ms/step - loss: 10388.9873 - r_square: -0.0104 - val_loss: 8968.7158 - val_r_square: -0.0445\n",
      "Epoch 370/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10386.6885 - r_square: -0.0102 - val_loss: 8836.2100 - val_r_square: -0.0291\n",
      "Epoch 371/400\n",
      "15/15 [==============================] - 1s 86ms/step - loss: 10434.2051 - r_square: -0.0148 - val_loss: 9580.2490 - val_r_square: -0.1157\n",
      "Epoch 372/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10302.2588 - r_square: -0.0019 - val_loss: 8972.7412 - val_r_square: -0.0450\n",
      "Epoch 373/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10386.1289 - r_square: -0.0101 - val_loss: 8820.8389 - val_r_square: -0.0273\n",
      "Epoch 374/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10420.0898 - r_square: -0.0134 - val_loss: 9313.6299 - val_r_square: -0.0847\n",
      "Epoch 375/400\n",
      "15/15 [==============================] - 1s 99ms/step - loss: 10305.5352 - r_square: -0.0023 - val_loss: 9613.9473 - val_r_square: -0.1197\n",
      "Epoch 376/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10324.2910 - r_square: -0.0041 - val_loss: 9005.2520 - val_r_square: -0.0488\n",
      "Epoch 377/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10466.5576 - r_square: -0.0179 - val_loss: 9802.2559 - val_r_square: -0.1416\n",
      "Epoch 378/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10310.7275 - r_square: -0.0028 - val_loss: 9186.2510 - val_r_square: -0.0699\n",
      "Epoch 379/400\n",
      "15/15 [==============================] - 1s 86ms/step - loss: 10375.1055 - r_square: -0.0090 - val_loss: 9399.9277 - val_r_square: -0.0947\n",
      "Epoch 380/400\n",
      "15/15 [==============================] - 1s 95ms/step - loss: 10323.2705 - r_square: -0.0040 - val_loss: 9678.0029 - val_r_square: -0.1271\n",
      "Epoch 381/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10306.4600 - r_square: -0.0024 - val_loss: 8904.5527 - val_r_square: -0.0371\n",
      "Epoch 382/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10377.9219 - r_square: -0.0093 - val_loss: 9196.5518 - val_r_square: -0.0711\n",
      "Epoch 383/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10327.5156 - r_square: -0.0044 - val_loss: 8905.1797 - val_r_square: -0.0371\n",
      "Epoch 384/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10310.9453 - r_square: -0.0028 - val_loss: 9686.3057 - val_r_square: -0.1281\n",
      "Epoch 385/400\n",
      "15/15 [==============================] - 1s 94ms/step - loss: 10350.6816 - r_square: -0.0067 - val_loss: 9572.2002 - val_r_square: -0.1148\n",
      "Epoch 386/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10514.6348 - r_square: -0.0226 - val_loss: 9158.9688 - val_r_square: -0.0667\n",
      "Epoch 387/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10585.9648 - r_square: -0.0295 - val_loss: 8643.9873 - val_r_square: -0.0067\n",
      "Epoch 388/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10375.8965 - r_square: -0.0091 - val_loss: 9414.7012 - val_r_square: -0.0965\n",
      "Epoch 389/400\n",
      "15/15 [==============================] - 1s 81ms/step - loss: 10314.9854 - r_square: -0.0032 - val_loss: 9539.5039 - val_r_square: -0.1110\n",
      "Epoch 390/400\n",
      "15/15 [==============================] - 1s 94ms/step - loss: 10318.9990 - r_square: -0.0036 - val_loss: 9538.5469 - val_r_square: -0.1109\n",
      "Epoch 391/400\n",
      "15/15 [==============================] - 1s 86ms/step - loss: 10318.1670 - r_square: -0.0035 - val_loss: 9275.4443 - val_r_square: -0.0802\n",
      "Epoch 392/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10351.6006 - r_square: -0.0067 - val_loss: 10058.4668 - val_r_square: -0.1714\n",
      "Epoch 393/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10287.9248 - r_square: -5.4646e-04 - val_loss: 8876.5205 - val_r_square: -0.0338\n",
      "Epoch 394/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10399.6729 - r_square: -0.0114 - val_loss: 9321.1504 - val_r_square: -0.0856\n",
      "Epoch 395/400\n",
      "15/15 [==============================] - 2s 108ms/step - loss: 10507.2100 - r_square: -0.0219 - val_loss: 10205.0459 - val_r_square: -0.1885\n",
      "Epoch 396/400\n",
      "15/15 [==============================] - 1s 87ms/step - loss: 10538.0996 - r_square: -0.0249 - val_loss: 9723.7256 - val_r_square: -0.1325\n",
      "Epoch 397/400\n",
      "15/15 [==============================] - 1s 84ms/step - loss: 10813.3164 - r_square: -0.0516 - val_loss: 8750.3975 - val_r_square: -0.0191\n",
      "Epoch 398/400\n",
      "15/15 [==============================] - 1s 83ms/step - loss: 10424.8154 - r_square: -0.0139 - val_loss: 8858.6846 - val_r_square: -0.0317\n",
      "Epoch 399/400\n",
      "15/15 [==============================] - 1s 82ms/step - loss: 10647.4707 - r_square: -0.0355 - val_loss: 9600.2354 - val_r_square: -0.1181\n",
      "Epoch 400/400\n",
      "15/15 [==============================] - 1s 98ms/step - loss: 10413.9824 - r_square: -0.0128 - val_loss: 10059.7041 - val_r_square: -0.1716\n"
     ]
    }
   ],
   "source": [
    "history_lstm2 = model_lstm2.fit(X_train_rs, y_train1, batch_size=64 , epochs=400, validation_data=(X_val_rs, y_val1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 364,
   "id": "5bb1a896-f6f4-4d09-8e5c-0cdb92721871",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 3s 21ms/step\n",
      "102.05035054944335\n"
     ]
    }
   ],
   "source": [
    "y_pred_lstm2_tr = model_lstm2.predict(X_train_rs)\n",
    "lstm2_rmse_train = np.sqrt(mean_squared_error(y_train, y_pred_lstm2_tr))\n",
    "print(lstm2_rmse_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 354,
   "id": "e61d52a7-029b-4bca-9a4a-229bc52b2bbd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history_lstm2)\n",
    "plt.savefig('PlotLoss_lstm2.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0bde9f5d-3b7c-40fe-a45e-63c704b3cc6c",
   "metadata": {
    "tags": []
   },
   "source": [
    "## ---------- Bidirectional LSTM ------------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 309,
   "id": "4d6aef06-360d-4962-8c32-23053cd0cfd7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "3/3 [==============================] - 9s 507ms/step - loss: 391262.1250 - r_square: -37.0525 - val_loss: 390632.4062 - val_r_square: -44.4939\n",
      "Epoch 2/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 620576.4375 - r_square: -59.3542 - val_loss: 141634.3906 - val_r_square: -15.4950\n",
      "Epoch 3/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 227420.7656 - r_square: -21.1179 - val_loss: 395528.8438 - val_r_square: -45.0641\n",
      "Epoch 4/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 371431.9375 - r_square: -35.1237 - val_loss: 416876.5000 - val_r_square: -47.5503\n",
      "Epoch 5/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 385962.4688 - r_square: -36.5375 - val_loss: 420849.2812 - val_r_square: -48.0130\n",
      "Epoch 6/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 388963.3125 - r_square: -36.8288 - val_loss: 421996.9688 - val_r_square: -48.1467\n",
      "Epoch 7/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 389804.7188 - r_square: -36.9104 - val_loss: 422129.0938 - val_r_square: -48.1621\n",
      "Epoch 8/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 389683.2188 - r_square: -36.8986 - val_loss: 421294.8750 - val_r_square: -48.0649\n",
      "Epoch 9/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 388332.7188 - r_square: -36.7676 - val_loss: 418497.9062 - val_r_square: -47.7392\n",
      "Epoch 10/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 384443.5000 - r_square: -36.3892 - val_loss: 411364.2188 - val_r_square: -46.9084\n",
      "Epoch 11/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 375363.8438 - r_square: -35.5062 - val_loss: 395899.8750 - val_r_square: -45.1074\n",
      "Epoch 12/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 355994.6562 - r_square: -33.6225 - val_loss: 362606.9375 - val_r_square: -41.2300\n",
      "Epoch 13/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 314601.1875 - r_square: -29.5965 - val_loss: 289762.8750 - val_r_square: -32.7464\n",
      "Epoch 14/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 223006.0156 - r_square: -20.6885 - val_loss: 135169.7344 - val_r_square: -14.7422\n",
      "Epoch 15/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 78434.2188 - r_square: -6.6282 - val_loss: 68226.0781 - val_r_square: -6.9458\n",
      "Epoch 16/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 118349.5859 - r_square: -10.5101 - val_loss: 39874.5820 - val_r_square: -3.6439\n",
      "Epoch 17/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 46404.3008 - r_square: -3.5131 - val_loss: 63664.7266 - val_r_square: -6.4145\n",
      "Epoch 18/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 61563.4453 - r_square: -4.9874 - val_loss: 82767.8750 - val_r_square: -8.6393\n",
      "Epoch 19/500\n",
      "3/3 [==============================] - 0s 66ms/step - loss: 62388.6445 - r_square: -5.0676 - val_loss: 46460.8945 - val_r_square: -4.4109\n",
      "Epoch 20/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 30354.3223 - r_square: -1.9521 - val_loss: 20657.3457 - val_r_square: -1.4058\n",
      "Epoch 21/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 33433.6250 - r_square: -2.2516 - val_loss: 25698.3887 - val_r_square: -1.9929\n",
      "Epoch 22/500\n",
      "3/3 [==============================] - 0s 63ms/step - loss: 27458.7715 - r_square: -1.6705 - val_loss: 19770.1465 - val_r_square: -1.3025\n",
      "Epoch 23/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 20397.4688 - r_square: -0.9838 - val_loss: 30762.7051 - val_r_square: -2.5827\n",
      "Epoch 24/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 24835.0820 - r_square: -1.4153 - val_loss: 24625.2441 - val_r_square: -1.8679\n",
      "Epoch 25/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 18409.4023 - r_square: -0.7904 - val_loss: 15078.7412 - val_r_square: -0.7561\n",
      "Epoch 26/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 17373.7012 - r_square: -0.6897 - val_loss: 15515.4229 - val_r_square: -0.8070\n",
      "Epoch 27/500\n",
      "3/3 [==============================] - 0s 65ms/step - loss: 17160.7344 - r_square: -0.6690 - val_loss: 14340.7354 - val_r_square: -0.6702\n",
      "Epoch 28/500\n",
      "3/3 [==============================] - 0s 66ms/step - loss: 13958.1973 - r_square: -0.3575 - val_loss: 17768.5742 - val_r_square: -1.0694\n",
      "Epoch 29/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 14929.5430 - r_square: -0.4520 - val_loss: 16624.3926 - val_r_square: -0.9361\n",
      "Epoch 30/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 13463.2031 - r_square: -0.3094 - val_loss: 13193.0381 - val_r_square: -0.5365\n",
      "Epoch 31/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 12923.4922 - r_square: -0.2569 - val_loss: 12565.0439 - val_r_square: -0.4634\n",
      "Epoch 32/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 12634.4443 - r_square: -0.2288 - val_loss: 12590.2900 - val_r_square: -0.4663\n",
      "Epoch 33/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 11924.5303 - r_square: -0.1597 - val_loss: 13948.1943 - val_r_square: -0.6244\n",
      "Epoch 34/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 11748.2959 - r_square: -0.1426 - val_loss: 12949.7568 - val_r_square: -0.5082\n",
      "Epoch 35/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 11051.0498 - r_square: -0.0748 - val_loss: 11549.6689 - val_r_square: -0.3451\n",
      "Epoch 36/500\n",
      "3/3 [==============================] - 0s 65ms/step - loss: 10766.9004 - r_square: -0.0471 - val_loss: 11147.0850 - val_r_square: -0.2982\n",
      "Epoch 37/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 10511.2754 - r_square: -0.0223 - val_loss: 11158.5986 - val_r_square: -0.2996\n",
      "Epoch 38/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 10156.6836 - r_square: 0.0122 - val_loss: 11685.6299 - val_r_square: -0.3609\n",
      "Epoch 39/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 9987.5068 - r_square: 0.0287 - val_loss: 11044.6455 - val_r_square: -0.2863\n",
      "Epoch 40/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 9649.1660 - r_square: 0.0616 - val_loss: 10170.6035 - val_r_square: -0.1845\n",
      "Epoch 41/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 9522.8018 - r_square: 0.0739 - val_loss: 9968.7295 - val_r_square: -0.1610\n",
      "Epoch 42/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 9134.9609 - r_square: 0.1116 - val_loss: 10530.4453 - val_r_square: -0.2264\n",
      "Epoch 43/500\n",
      "3/3 [==============================] - 0s 83ms/step - loss: 9059.0439 - r_square: 0.1190 - val_loss: 10280.3691 - val_r_square: -0.1973\n",
      "Epoch 44/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 8774.0576 - r_square: 0.1467 - val_loss: 9416.8867 - val_r_square: -0.0967\n",
      "Epoch 45/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 8571.2520 - r_square: 0.1664 - val_loss: 9007.2891 - val_r_square: -0.0490\n",
      "Epoch 46/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 8428.9883 - r_square: 0.1802 - val_loss: 9080.7695 - val_r_square: -0.0576\n",
      "Epoch 47/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 8200.4844 - r_square: 0.2025 - val_loss: 9356.8848 - val_r_square: -0.0897\n",
      "Epoch 48/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 8091.5493 - r_square: 0.2131 - val_loss: 8928.1611 - val_r_square: -0.0398\n",
      "Epoch 49/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 7888.4751 - r_square: 0.2328 - val_loss: 8331.8008 - val_r_square: 0.0297\n",
      "Epoch 50/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 7813.7192 - r_square: 0.2401 - val_loss: 8379.0898 - val_r_square: 0.0242\n",
      "Epoch 51/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 7598.5195 - r_square: 0.2610 - val_loss: 8365.7715 - val_r_square: 0.0257\n",
      "Epoch 52/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 7464.8374 - r_square: 0.2740 - val_loss: 8192.7744 - val_r_square: 0.0459\n",
      "Epoch 53/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 7345.7446 - r_square: 0.2856 - val_loss: 8044.5234 - val_r_square: 0.0631\n",
      "Epoch 54/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 7225.2393 - r_square: 0.2973 - val_loss: 7992.2603 - val_r_square: 0.0692\n",
      "Epoch 55/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 7114.8516 - r_square: 0.3080 - val_loss: 7803.3892 - val_r_square: 0.0912\n",
      "Epoch 56/500\n",
      "3/3 [==============================] - 1s 336ms/step - loss: 7016.2310 - r_square: 0.3176 - val_loss: 7670.3711 - val_r_square: 0.1067\n",
      "Epoch 57/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 6931.0352 - r_square: 0.3259 - val_loss: 7505.6035 - val_r_square: 0.1259\n",
      "Epoch 58/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 6823.8765 - r_square: 0.3363 - val_loss: 7590.4980 - val_r_square: 0.1160\n",
      "Epoch 59/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 6744.5684 - r_square: 0.3441 - val_loss: 7548.9614 - val_r_square: 0.1208\n",
      "Epoch 60/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 6661.4858 - r_square: 0.3521 - val_loss: 7256.1665 - val_r_square: 0.1549\n",
      "Epoch 61/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 6572.8164 - r_square: 0.3608 - val_loss: 7135.8945 - val_r_square: 0.1689\n",
      "Epoch 62/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 6503.7056 - r_square: 0.3675 - val_loss: 7114.1001 - val_r_square: 0.1715\n",
      "Epoch 63/500\n",
      "3/3 [==============================] - 0s 66ms/step - loss: 6446.8560 - r_square: 0.3730 - val_loss: 7004.5879 - val_r_square: 0.1842\n",
      "Epoch 64/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 6352.0435 - r_square: 0.3822 - val_loss: 6833.5586 - val_r_square: 0.2041\n",
      "Epoch 65/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 6288.2451 - r_square: 0.3884 - val_loss: 6822.9946 - val_r_square: 0.2054\n",
      "Epoch 66/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 6221.6211 - r_square: 0.3949 - val_loss: 6820.6909 - val_r_square: 0.2056\n",
      "Epoch 67/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 6166.4326 - r_square: 0.4003 - val_loss: 6658.4712 - val_r_square: 0.2245\n",
      "Epoch 68/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 6089.4595 - r_square: 0.4078 - val_loss: 6393.8374 - val_r_square: 0.2554\n",
      "Epoch 69/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 6034.1909 - r_square: 0.4131 - val_loss: 6662.8765 - val_r_square: 0.2240\n",
      "Epoch 70/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 5975.4639 - r_square: 0.4189 - val_loss: 6627.4175 - val_r_square: 0.2282\n",
      "Epoch 71/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 5921.0322 - r_square: 0.4241 - val_loss: 6219.1128 - val_r_square: 0.2757\n",
      "Epoch 72/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 5836.2080 - r_square: 0.4324 - val_loss: 6422.8916 - val_r_square: 0.2520\n",
      "Epoch 73/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 5793.3794 - r_square: 0.4366 - val_loss: 6226.0371 - val_r_square: 0.2749\n",
      "Epoch 74/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 5706.5723 - r_square: 0.4450 - val_loss: 5925.6157 - val_r_square: 0.3099\n",
      "Epoch 75/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 5670.5767 - r_square: 0.4485 - val_loss: 6269.5332 - val_r_square: 0.2698\n",
      "Epoch 76/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 5581.5679 - r_square: 0.4572 - val_loss: 5827.9521 - val_r_square: 0.3213\n",
      "Epoch 77/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 5514.5615 - r_square: 0.4637 - val_loss: 6019.6309 - val_r_square: 0.2989\n",
      "Epoch 78/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 5461.1611 - r_square: 0.4689 - val_loss: 5708.5713 - val_r_square: 0.3352\n",
      "Epoch 79/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 5457.8965 - r_square: 0.4692 - val_loss: 5889.6987 - val_r_square: 0.3141\n",
      "Epoch 80/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 5407.5317 - r_square: 0.4741 - val_loss: 5584.2861 - val_r_square: 0.3496\n",
      "Epoch 81/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 5340.0166 - r_square: 0.4807 - val_loss: 5506.1870 - val_r_square: 0.3587\n",
      "Epoch 82/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 5246.9082 - r_square: 0.4897 - val_loss: 5598.7256 - val_r_square: 0.3480\n",
      "Epoch 83/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 5206.8516 - r_square: 0.4936 - val_loss: 5341.4868 - val_r_square: 0.3779\n",
      "Epoch 84/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 5103.4092 - r_square: 0.5037 - val_loss: 5541.4067 - val_r_square: 0.3546\n",
      "Epoch 85/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 5056.6299 - r_square: 0.5082 - val_loss: 5188.8096 - val_r_square: 0.3957\n",
      "Epoch 86/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 5006.1055 - r_square: 0.5131 - val_loss: 5141.1948 - val_r_square: 0.4012\n",
      "Epoch 87/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 5010.5361 - r_square: 0.5127 - val_loss: 5298.4678 - val_r_square: 0.3829\n",
      "Epoch 88/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 4822.8901 - r_square: 0.5309 - val_loss: 5085.8364 - val_r_square: 0.4077\n",
      "Epoch 89/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4710.2998 - r_square: 0.5419 - val_loss: 4790.9204 - val_r_square: 0.4420\n",
      "Epoch 90/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 4685.7495 - r_square: 0.5443 - val_loss: 4994.5767 - val_r_square: 0.4183\n",
      "Epoch 91/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 4608.8179 - r_square: 0.5518 - val_loss: 4632.8237 - val_r_square: 0.4605\n",
      "Epoch 92/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 4581.5015 - r_square: 0.5544 - val_loss: 5191.7739 - val_r_square: 0.3954\n",
      "Epoch 93/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 4588.7422 - r_square: 0.5537 - val_loss: 4654.2329 - val_r_square: 0.4580\n",
      "Epoch 94/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 4539.3945 - r_square: 0.5585 - val_loss: 4584.0566 - val_r_square: 0.4661\n",
      "Epoch 95/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 4618.2065 - r_square: 0.5509 - val_loss: 5993.6606 - val_r_square: 0.3020\n",
      "Epoch 96/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 4829.0205 - r_square: 0.5303 - val_loss: 4348.6792 - val_r_square: 0.4935\n",
      "Epoch 97/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4498.8257 - r_square: 0.5625 - val_loss: 4946.9424 - val_r_square: 0.4239\n",
      "Epoch 98/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 4463.5708 - r_square: 0.5659 - val_loss: 4618.6353 - val_r_square: 0.4621\n",
      "Epoch 99/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 4532.5259 - r_square: 0.5592 - val_loss: 4311.3813 - val_r_square: 0.4979\n",
      "Epoch 100/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 4488.2080 - r_square: 0.5635 - val_loss: 5339.9434 - val_r_square: 0.3781\n",
      "Epoch 101/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 4546.1562 - r_square: 0.5579 - val_loss: 4341.3940 - val_r_square: 0.4944\n",
      "Epoch 102/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4703.6514 - r_square: 0.5425 - val_loss: 5103.2573 - val_r_square: 0.4057\n",
      "Epoch 103/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 4409.0645 - r_square: 0.5712 - val_loss: 4375.4482 - val_r_square: 0.4904\n",
      "Epoch 104/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 4456.2773 - r_square: 0.5666 - val_loss: 4420.3433 - val_r_square: 0.4852\n",
      "Epoch 105/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 4391.5015 - r_square: 0.5729 - val_loss: 4751.8086 - val_r_square: 0.4466\n",
      "Epoch 106/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4341.3779 - r_square: 0.5778 - val_loss: 4319.4438 - val_r_square: 0.4969\n",
      "Epoch 107/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4318.6567 - r_square: 0.5800 - val_loss: 4592.8062 - val_r_square: 0.4651\n",
      "Epoch 108/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 4227.8862 - r_square: 0.5888 - val_loss: 4414.2451 - val_r_square: 0.4859\n",
      "Epoch 109/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4238.3018 - r_square: 0.5878 - val_loss: 4325.6270 - val_r_square: 0.4962\n",
      "Epoch 110/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 4262.0103 - r_square: 0.5855 - val_loss: 5427.2627 - val_r_square: 0.3679\n",
      "Epoch 111/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 4452.5415 - r_square: 0.5670 - val_loss: 4189.1558 - val_r_square: 0.5121\n",
      "Epoch 112/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 4210.9150 - r_square: 0.5905 - val_loss: 5301.2095 - val_r_square: 0.3826\n",
      "Epoch 113/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4322.0674 - r_square: 0.5797 - val_loss: 4144.9736 - val_r_square: 0.5173\n",
      "Epoch 114/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 4360.3413 - r_square: 0.5759 - val_loss: 4382.9976 - val_r_square: 0.4895\n",
      "Epoch 115/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 4124.8667 - r_square: 0.5988 - val_loss: 4298.2700 - val_r_square: 0.4994\n",
      "Epoch 116/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 4085.0676 - r_square: 0.6027 - val_loss: 4481.3589 - val_r_square: 0.4781\n",
      "Epoch 117/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 4103.8066 - r_square: 0.6009 - val_loss: 4197.4644 - val_r_square: 0.5112\n",
      "Epoch 118/500\n",
      "3/3 [==============================] - 0s 82ms/step - loss: 4103.9009 - r_square: 0.6009 - val_loss: 4365.1558 - val_r_square: 0.4916\n",
      "Epoch 119/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 4073.0403 - r_square: 0.6039 - val_loss: 4383.6816 - val_r_square: 0.4895\n",
      "Epoch 120/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 4074.1057 - r_square: 0.6038 - val_loss: 4119.1626 - val_r_square: 0.5203\n",
      "Epoch 121/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 4092.3701 - r_square: 0.6020 - val_loss: 4611.6860 - val_r_square: 0.4629\n",
      "Epoch 122/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 4054.1509 - r_square: 0.6057 - val_loss: 4034.4836 - val_r_square: 0.5301\n",
      "Epoch 123/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 4051.1746 - r_square: 0.6060 - val_loss: 4666.6826 - val_r_square: 0.4565\n",
      "Epoch 124/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 4071.9380 - r_square: 0.6040 - val_loss: 4175.2480 - val_r_square: 0.5137\n",
      "Epoch 125/500\n",
      "3/3 [==============================] - 0s 82ms/step - loss: 4062.1414 - r_square: 0.6049 - val_loss: 4045.5203 - val_r_square: 0.5288\n",
      "Epoch 126/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 4028.0239 - r_square: 0.6083 - val_loss: 4521.5376 - val_r_square: 0.4734\n",
      "Epoch 127/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4013.8564 - r_square: 0.6096 - val_loss: 4016.1936 - val_r_square: 0.5323\n",
      "Epoch 128/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 4217.9126 - r_square: 0.5898 - val_loss: 4477.7412 - val_r_square: 0.4785\n",
      "Epoch 129/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 4091.6492 - r_square: 0.6021 - val_loss: 4583.3296 - val_r_square: 0.4662\n",
      "Epoch 130/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 4108.8330 - r_square: 0.6004 - val_loss: 4001.6226 - val_r_square: 0.5340\n",
      "Epoch 131/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4120.0190 - r_square: 0.5993 - val_loss: 4878.2856 - val_r_square: 0.4319\n",
      "Epoch 132/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 4022.7739 - r_square: 0.6088 - val_loss: 4003.8074 - val_r_square: 0.5337\n",
      "Epoch 133/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4101.2446 - r_square: 0.6011 - val_loss: 3990.1824 - val_r_square: 0.5353\n",
      "Epoch 134/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4046.8789 - r_square: 0.6064 - val_loss: 4593.3730 - val_r_square: 0.4650\n",
      "Epoch 135/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3997.6135 - r_square: 0.6112 - val_loss: 3982.7937 - val_r_square: 0.5362\n",
      "Epoch 136/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 4284.5381 - r_square: 0.5833 - val_loss: 4459.3848 - val_r_square: 0.4807\n",
      "Epoch 137/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 4012.7341 - r_square: 0.6097 - val_loss: 4389.1724 - val_r_square: 0.4888\n",
      "Epoch 138/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 3942.2324 - r_square: 0.6166 - val_loss: 3974.2783 - val_r_square: 0.5371\n",
      "Epoch 139/500\n",
      "3/3 [==============================] - 0s 81ms/step - loss: 3910.0703 - r_square: 0.6197 - val_loss: 3990.5100 - val_r_square: 0.5353\n",
      "Epoch 140/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3923.3071 - r_square: 0.6184 - val_loss: 4339.4985 - val_r_square: 0.4946\n",
      "Epoch 141/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3891.9863 - r_square: 0.6215 - val_loss: 4303.3276 - val_r_square: 0.4988\n",
      "Epoch 142/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 3900.3569 - r_square: 0.6207 - val_loss: 3955.2000 - val_r_square: 0.5394\n",
      "Epoch 143/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3929.3489 - r_square: 0.6178 - val_loss: 4282.7114 - val_r_square: 0.5012\n",
      "Epoch 144/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3854.1621 - r_square: 0.6252 - val_loss: 4176.2441 - val_r_square: 0.5136\n",
      "Epoch 145/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3823.0979 - r_square: 0.6282 - val_loss: 4067.5361 - val_r_square: 0.5263\n",
      "Epoch 146/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3806.5813 - r_square: 0.6298 - val_loss: 4520.5127 - val_r_square: 0.4735\n",
      "Epoch 147/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3898.1440 - r_square: 0.6209 - val_loss: 4158.6050 - val_r_square: 0.5157\n",
      "Epoch 148/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3792.5476 - r_square: 0.6312 - val_loss: 4410.2261 - val_r_square: 0.4864\n",
      "Epoch 149/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3803.8162 - r_square: 0.6301 - val_loss: 3942.9824 - val_r_square: 0.5408\n",
      "Epoch 150/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3868.2336 - r_square: 0.6238 - val_loss: 3956.5991 - val_r_square: 0.5392\n",
      "Epoch 151/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3781.5632 - r_square: 0.6322 - val_loss: 4303.0444 - val_r_square: 0.4989\n",
      "Epoch 152/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3786.5557 - r_square: 0.6317 - val_loss: 4016.4463 - val_r_square: 0.5322\n",
      "Epoch 153/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3764.3955 - r_square: 0.6339 - val_loss: 4122.5586 - val_r_square: 0.5199\n",
      "Epoch 154/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3732.1106 - r_square: 0.6370 - val_loss: 4228.2446 - val_r_square: 0.5076\n",
      "Epoch 155/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 3754.9583 - r_square: 0.6348 - val_loss: 3907.5605 - val_r_square: 0.5449\n",
      "Epoch 156/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3707.7253 - r_square: 0.6394 - val_loss: 4445.0791 - val_r_square: 0.4823\n",
      "Epoch 157/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3781.1570 - r_square: 0.6323 - val_loss: 4047.0518 - val_r_square: 0.5287\n",
      "Epoch 158/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3737.7920 - r_square: 0.6365 - val_loss: 3894.8118 - val_r_square: 0.5464\n",
      "Epoch 159/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3693.8389 - r_square: 0.6408 - val_loss: 4103.0908 - val_r_square: 0.5221\n",
      "Epoch 160/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3711.1682 - r_square: 0.6391 - val_loss: 3956.8088 - val_r_square: 0.5392\n",
      "Epoch 161/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3786.9221 - r_square: 0.6317 - val_loss: 3849.8015 - val_r_square: 0.5516\n",
      "Epoch 162/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 3807.1235 - r_square: 0.6297 - val_loss: 4006.7114 - val_r_square: 0.5334\n",
      "Epoch 163/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3811.6223 - r_square: 0.6293 - val_loss: 4435.5425 - val_r_square: 0.4834\n",
      "Epoch 164/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3732.9290 - r_square: 0.6370 - val_loss: 3792.8611 - val_r_square: 0.5583\n",
      "Epoch 165/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3779.4067 - r_square: 0.6324 - val_loss: 3945.0466 - val_r_square: 0.5406\n",
      "Epoch 166/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3724.8640 - r_square: 0.6377 - val_loss: 4407.0688 - val_r_square: 0.4867\n",
      "Epoch 167/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3778.4956 - r_square: 0.6325 - val_loss: 3857.8088 - val_r_square: 0.5507\n",
      "Epoch 168/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3899.7222 - r_square: 0.6207 - val_loss: 4387.5269 - val_r_square: 0.4890\n",
      "Epoch 169/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3732.5398 - r_square: 0.6370 - val_loss: 4123.0894 - val_r_square: 0.5198\n",
      "Epoch 170/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3656.0132 - r_square: 0.6444 - val_loss: 3910.0864 - val_r_square: 0.5446\n",
      "Epoch 171/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3665.7417 - r_square: 0.6435 - val_loss: 4624.4517 - val_r_square: 0.4614\n",
      "Epoch 172/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3711.4885 - r_square: 0.6390 - val_loss: 3802.6228 - val_r_square: 0.5571\n",
      "Epoch 173/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3698.3333 - r_square: 0.6403 - val_loss: 3947.4353 - val_r_square: 0.5403\n",
      "Epoch 174/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3619.3152 - r_square: 0.6480 - val_loss: 4149.0220 - val_r_square: 0.5168\n",
      "Epoch 175/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3628.2495 - r_square: 0.6471 - val_loss: 4006.3865 - val_r_square: 0.5334\n",
      "Epoch 176/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3636.9683 - r_square: 0.6463 - val_loss: 3796.2285 - val_r_square: 0.5579\n",
      "Epoch 177/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3745.3865 - r_square: 0.6357 - val_loss: 4105.3911 - val_r_square: 0.5219\n",
      "Epoch 178/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3660.8918 - r_square: 0.6440 - val_loss: 4243.8481 - val_r_square: 0.5058\n",
      "Epoch 179/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3656.2830 - r_square: 0.6444 - val_loss: 3828.1792 - val_r_square: 0.5542\n",
      "Epoch 180/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 3665.1592 - r_square: 0.6435 - val_loss: 3908.5483 - val_r_square: 0.5448\n",
      "Epoch 181/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3583.6277 - r_square: 0.6515 - val_loss: 4373.6685 - val_r_square: 0.4906\n",
      "Epoch 182/500\n",
      "3/3 [==============================] - 0s 85ms/step - loss: 3702.1086 - r_square: 0.6400 - val_loss: 3930.8882 - val_r_square: 0.5422\n",
      "Epoch 183/500\n",
      "3/3 [==============================] - 0s 100ms/step - loss: 3597.7939 - r_square: 0.6501 - val_loss: 3890.7876 - val_r_square: 0.5469\n",
      "Epoch 184/500\n",
      "3/3 [==============================] - 0s 87ms/step - loss: 3579.0437 - r_square: 0.6519 - val_loss: 3923.8689 - val_r_square: 0.5430\n",
      "Epoch 185/500\n",
      "3/3 [==============================] - 0s 84ms/step - loss: 3568.1072 - r_square: 0.6530 - val_loss: 3794.0557 - val_r_square: 0.5581\n",
      "Epoch 186/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3687.8286 - r_square: 0.6413 - val_loss: 3757.4214 - val_r_square: 0.5624\n",
      "Epoch 187/500\n",
      "3/3 [==============================] - 0s 81ms/step - loss: 3669.3203 - r_square: 0.6431 - val_loss: 4242.8325 - val_r_square: 0.5059\n",
      "Epoch 188/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3609.6489 - r_square: 0.6489 - val_loss: 4114.8989 - val_r_square: 0.5208\n",
      "Epoch 189/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3568.4124 - r_square: 0.6530 - val_loss: 3844.7451 - val_r_square: 0.5522\n",
      "Epoch 190/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3572.9939 - r_square: 0.6525 - val_loss: 4314.0518 - val_r_square: 0.4976\n",
      "Epoch 191/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3667.5417 - r_square: 0.6433 - val_loss: 4002.1877 - val_r_square: 0.5339\n",
      "Epoch 192/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3872.3103 - r_square: 0.6234 - val_loss: 3752.8535 - val_r_square: 0.5629\n",
      "Epoch 193/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3887.4180 - r_square: 0.6219 - val_loss: 4719.8315 - val_r_square: 0.4503\n",
      "Epoch 194/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3750.3728 - r_square: 0.6352 - val_loss: 4189.6143 - val_r_square: 0.5121\n",
      "Epoch 195/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3690.3818 - r_square: 0.6411 - val_loss: 3961.5833 - val_r_square: 0.5386\n",
      "Epoch 196/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3934.6160 - r_square: 0.6173 - val_loss: 3937.6611 - val_r_square: 0.5414\n",
      "Epoch 197/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3706.3647 - r_square: 0.6395 - val_loss: 5275.7915 - val_r_square: 0.3856\n",
      "Epoch 198/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3901.2883 - r_square: 0.6206 - val_loss: 3832.2854 - val_r_square: 0.5537\n",
      "Epoch 199/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3888.9597 - r_square: 0.6218 - val_loss: 3847.9070 - val_r_square: 0.5519\n",
      "Epoch 200/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3638.7097 - r_square: 0.6461 - val_loss: 4514.1230 - val_r_square: 0.4743\n",
      "Epoch 201/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3604.4221 - r_square: 0.6495 - val_loss: 3784.8572 - val_r_square: 0.5592\n",
      "Epoch 202/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 3673.6597 - r_square: 0.6427 - val_loss: 4089.5237 - val_r_square: 0.5237\n",
      "Epoch 203/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3591.6196 - r_square: 0.6507 - val_loss: 4539.8940 - val_r_square: 0.4713\n",
      "Epoch 204/500\n",
      "3/3 [==============================] - 0s 95ms/step - loss: 3687.5876 - r_square: 0.6414 - val_loss: 3822.6978 - val_r_square: 0.5548\n",
      "Epoch 205/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3856.6382 - r_square: 0.6249 - val_loss: 3905.1125 - val_r_square: 0.5452\n",
      "Epoch 206/500\n",
      "3/3 [==============================] - 0s 97ms/step - loss: 3633.4934 - r_square: 0.6466 - val_loss: 4241.8276 - val_r_square: 0.5060\n",
      "Epoch 207/500\n",
      "3/3 [==============================] - 0s 83ms/step - loss: 3528.7773 - r_square: 0.6568 - val_loss: 3868.8972 - val_r_square: 0.5494\n",
      "Epoch 208/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3514.0693 - r_square: 0.6582 - val_loss: 3702.3667 - val_r_square: 0.5688\n",
      "Epoch 209/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 3599.4163 - r_square: 0.6499 - val_loss: 3653.1426 - val_r_square: 0.5745\n",
      "Epoch 210/500\n",
      "3/3 [==============================] - 0s 95ms/step - loss: 3547.3777 - r_square: 0.6550 - val_loss: 4675.8408 - val_r_square: 0.4554\n",
      "Epoch 211/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3671.1050 - r_square: 0.6430 - val_loss: 3911.6245 - val_r_square: 0.5444\n",
      "Epoch 212/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3537.0054 - r_square: 0.6560 - val_loss: 3689.1868 - val_r_square: 0.5703\n",
      "Epoch 213/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3545.6162 - r_square: 0.6552 - val_loss: 3894.7288 - val_r_square: 0.5464\n",
      "Epoch 214/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3484.6873 - r_square: 0.6611 - val_loss: 4180.2588 - val_r_square: 0.5132\n",
      "Epoch 215/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3552.5190 - r_square: 0.6545 - val_loss: 3766.8564 - val_r_square: 0.5613\n",
      "Epoch 216/500\n",
      "3/3 [==============================] - 1s 290ms/step - loss: 3701.9880 - r_square: 0.6400 - val_loss: 3714.9114 - val_r_square: 0.5674\n",
      "Epoch 217/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3481.6316 - r_square: 0.6614 - val_loss: 5012.0615 - val_r_square: 0.4163\n",
      "Epoch 218/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 3760.3474 - r_square: 0.6343 - val_loss: 3685.2944 - val_r_square: 0.5708\n",
      "Epoch 219/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 3728.5085 - r_square: 0.6374 - val_loss: 3559.2017 - val_r_square: 0.5855\n",
      "Epoch 220/500\n",
      "3/3 [==============================] - 0s 88ms/step - loss: 3816.2539 - r_square: 0.6289 - val_loss: 4745.0640 - val_r_square: 0.4474\n",
      "Epoch 221/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3584.1282 - r_square: 0.6514 - val_loss: 3836.4282 - val_r_square: 0.5532\n",
      "Epoch 222/500\n",
      "3/3 [==============================] - 0s 83ms/step - loss: 3715.1250 - r_square: 0.6387 - val_loss: 3640.5747 - val_r_square: 0.5760\n",
      "Epoch 223/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3736.0361 - r_square: 0.6367 - val_loss: 4233.8213 - val_r_square: 0.5069\n",
      "Epoch 224/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3659.0225 - r_square: 0.6441 - val_loss: 4552.2412 - val_r_square: 0.4698\n",
      "Epoch 225/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3792.1133 - r_square: 0.6312 - val_loss: 3774.4854 - val_r_square: 0.5604\n",
      "Epoch 226/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3598.5930 - r_square: 0.6500 - val_loss: 3916.6453 - val_r_square: 0.5439\n",
      "Epoch 227/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3487.6953 - r_square: 0.6608 - val_loss: 3891.1414 - val_r_square: 0.5468\n",
      "Epoch 228/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3445.1108 - r_square: 0.6649 - val_loss: 4412.4116 - val_r_square: 0.4861\n",
      "Epoch 229/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3505.3491 - r_square: 0.6591 - val_loss: 3829.7996 - val_r_square: 0.5540\n",
      "Epoch 230/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3692.9771 - r_square: 0.6408 - val_loss: 3896.2603 - val_r_square: 0.5462\n",
      "Epoch 231/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3592.6855 - r_square: 0.6506 - val_loss: 4817.2319 - val_r_square: 0.4390\n",
      "Epoch 232/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3690.2910 - r_square: 0.6411 - val_loss: 3644.4468 - val_r_square: 0.5756\n",
      "Epoch 233/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3484.5986 - r_square: 0.6611 - val_loss: 4084.9668 - val_r_square: 0.5243\n",
      "Epoch 234/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3561.4036 - r_square: 0.6536 - val_loss: 4051.3586 - val_r_square: 0.5282\n",
      "Epoch 235/500\n",
      "3/3 [==============================] - 0s 67ms/step - loss: 3399.3811 - r_square: 0.6694 - val_loss: 3710.9534 - val_r_square: 0.5678\n",
      "Epoch 236/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3397.8521 - r_square: 0.6695 - val_loss: 3772.8235 - val_r_square: 0.5606\n",
      "Epoch 237/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3372.9692 - r_square: 0.6720 - val_loss: 4168.4136 - val_r_square: 0.5145\n",
      "Epoch 238/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3369.8149 - r_square: 0.6723 - val_loss: 3667.9888 - val_r_square: 0.5728\n",
      "Epoch 239/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3577.5571 - r_square: 0.6521 - val_loss: 3691.2266 - val_r_square: 0.5701\n",
      "Epoch 240/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3563.4751 - r_square: 0.6534 - val_loss: 5117.7031 - val_r_square: 0.4040\n",
      "Epoch 241/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3702.5955 - r_square: 0.6399 - val_loss: 3679.3464 - val_r_square: 0.5715\n",
      "Epoch 242/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3501.9214 - r_square: 0.6594 - val_loss: 3614.8765 - val_r_square: 0.5790\n",
      "Epoch 243/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3630.6819 - r_square: 0.6469 - val_loss: 4183.6787 - val_r_square: 0.5128\n",
      "Epoch 244/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3469.0889 - r_square: 0.6626 - val_loss: 4132.4639 - val_r_square: 0.5187\n",
      "Epoch 245/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3522.8442 - r_square: 0.6574 - val_loss: 3653.0598 - val_r_square: 0.5746\n",
      "Epoch 246/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3546.2288 - r_square: 0.6551 - val_loss: 4143.9766 - val_r_square: 0.5174\n",
      "Epoch 247/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3546.5903 - r_square: 0.6551 - val_loss: 4254.4824 - val_r_square: 0.5045\n",
      "Epoch 248/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3541.6194 - r_square: 0.6556 - val_loss: 3866.6492 - val_r_square: 0.5497\n",
      "Epoch 249/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3515.4077 - r_square: 0.6581 - val_loss: 3839.4800 - val_r_square: 0.5528\n",
      "Epoch 250/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3411.9683 - r_square: 0.6682 - val_loss: 3725.8096 - val_r_square: 0.5661\n",
      "Epoch 251/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 3351.2969 - r_square: 0.6741 - val_loss: 3757.2080 - val_r_square: 0.5624\n",
      "Epoch 252/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3359.5178 - r_square: 0.6733 - val_loss: 3866.2410 - val_r_square: 0.5497\n",
      "Epoch 253/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3357.6167 - r_square: 0.6735 - val_loss: 3683.5828 - val_r_square: 0.5710\n",
      "Epoch 254/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3509.3684 - r_square: 0.6587 - val_loss: 3613.6326 - val_r_square: 0.5791\n",
      "Epoch 255/500\n",
      "3/3 [==============================] - 0s 84ms/step - loss: 3494.5283 - r_square: 0.6601 - val_loss: 3792.7505 - val_r_square: 0.5583\n",
      "Epoch 256/500\n",
      "3/3 [==============================] - 0s 81ms/step - loss: 3399.9231 - r_square: 0.6693 - val_loss: 4400.9722 - val_r_square: 0.4875\n",
      "Epoch 257/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3541.7959 - r_square: 0.6555 - val_loss: 3861.4036 - val_r_square: 0.5503\n",
      "Epoch 258/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3521.1855 - r_square: 0.6575 - val_loss: 3724.8420 - val_r_square: 0.5662\n",
      "Epoch 259/500\n",
      "3/3 [==============================] - 0s 84ms/step - loss: 3768.1982 - r_square: 0.6335 - val_loss: 3698.3489 - val_r_square: 0.5693\n",
      "Epoch 260/500\n",
      "3/3 [==============================] - 0s 89ms/step - loss: 3505.3132 - r_square: 0.6591 - val_loss: 5305.6216 - val_r_square: 0.3821\n",
      "Epoch 261/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3808.4973 - r_square: 0.6296 - val_loss: 3603.4150 - val_r_square: 0.5803\n",
      "Epoch 262/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3742.8677 - r_square: 0.6360 - val_loss: 3654.0364 - val_r_square: 0.5744\n",
      "Epoch 263/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3569.0364 - r_square: 0.6529 - val_loss: 4361.0176 - val_r_square: 0.4921\n",
      "Epoch 264/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 3644.8037 - r_square: 0.6455 - val_loss: 4290.4990 - val_r_square: 0.5003\n",
      "Epoch 265/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3419.4390 - r_square: 0.6674 - val_loss: 3553.5940 - val_r_square: 0.5861\n",
      "Epoch 266/500\n",
      "3/3 [==============================] - 0s 66ms/step - loss: 3375.8992 - r_square: 0.6717 - val_loss: 3690.3677 - val_r_square: 0.5702\n",
      "Epoch 267/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3437.7769 - r_square: 0.6657 - val_loss: 3892.9172 - val_r_square: 0.5466\n",
      "Epoch 268/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3333.4255 - r_square: 0.6758 - val_loss: 4000.5186 - val_r_square: 0.5341\n",
      "Epoch 269/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3291.0681 - r_square: 0.6799 - val_loss: 3543.8730 - val_r_square: 0.5873\n",
      "Epoch 270/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3321.5447 - r_square: 0.6770 - val_loss: 4280.0146 - val_r_square: 0.5015\n",
      "Epoch 271/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3527.2080 - r_square: 0.6570 - val_loss: 4707.2095 - val_r_square: 0.4518\n",
      "Epoch 272/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3746.6931 - r_square: 0.6356 - val_loss: 3622.2664 - val_r_square: 0.5781\n",
      "Epoch 273/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3661.1775 - r_square: 0.6439 - val_loss: 3613.2417 - val_r_square: 0.5792\n",
      "Epoch 274/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3408.2717 - r_square: 0.6685 - val_loss: 5010.8853 - val_r_square: 0.4164\n",
      "Epoch 275/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3609.1333 - r_square: 0.6490 - val_loss: 3612.0276 - val_r_square: 0.5793\n",
      "Epoch 276/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3739.1228 - r_square: 0.6363 - val_loss: 3462.4800 - val_r_square: 0.5968\n",
      "Epoch 277/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3501.7498 - r_square: 0.6594 - val_loss: 3632.3171 - val_r_square: 0.5770\n",
      "Epoch 278/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 3351.7986 - r_square: 0.6740 - val_loss: 3774.7585 - val_r_square: 0.5604\n",
      "Epoch 279/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3481.8782 - r_square: 0.6614 - val_loss: 3725.8123 - val_r_square: 0.5661\n",
      "Epoch 280/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3494.6250 - r_square: 0.6601 - val_loss: 4033.9001 - val_r_square: 0.5302\n",
      "Epoch 281/500\n",
      "3/3 [==============================] - 0s 81ms/step - loss: 3345.5906 - r_square: 0.6746 - val_loss: 4079.3845 - val_r_square: 0.5249\n",
      "Epoch 282/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3430.1611 - r_square: 0.6664 - val_loss: 3580.0063 - val_r_square: 0.5831\n",
      "Epoch 283/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3390.9021 - r_square: 0.6702 - val_loss: 3497.1941 - val_r_square: 0.5927\n",
      "Epoch 284/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3350.5813 - r_square: 0.6741 - val_loss: 4031.9995 - val_r_square: 0.5304\n",
      "Epoch 285/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3317.4707 - r_square: 0.6774 - val_loss: 3669.5039 - val_r_square: 0.5726\n",
      "Epoch 286/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3260.2854 - r_square: 0.6829 - val_loss: 3516.1499 - val_r_square: 0.5905\n",
      "Epoch 287/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3253.4133 - r_square: 0.6836 - val_loss: 4328.9092 - val_r_square: 0.4958\n",
      "Epoch 288/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3286.8130 - r_square: 0.6803 - val_loss: 3587.7500 - val_r_square: 0.5822\n",
      "Epoch 289/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3393.9695 - r_square: 0.6699 - val_loss: 3553.0872 - val_r_square: 0.5862\n",
      "Epoch 290/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3350.1646 - r_square: 0.6742 - val_loss: 5026.7007 - val_r_square: 0.4146\n",
      "Epoch 291/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3472.2366 - r_square: 0.6623 - val_loss: 3622.3792 - val_r_square: 0.5781\n",
      "Epoch 292/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3666.1721 - r_square: 0.6434 - val_loss: 3494.4927 - val_r_square: 0.5930\n",
      "Epoch 293/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3618.8076 - r_square: 0.6481 - val_loss: 4834.5967 - val_r_square: 0.4370\n",
      "Epoch 294/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 3768.9797 - r_square: 0.6334 - val_loss: 4855.3340 - val_r_square: 0.4345\n",
      "Epoch 295/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 3878.0569 - r_square: 0.6228 - val_loss: 4014.5930 - val_r_square: 0.5325\n",
      "Epoch 296/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 4059.3936 - r_square: 0.6052 - val_loss: 3753.8701 - val_r_square: 0.5628\n",
      "Epoch 297/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3630.9968 - r_square: 0.6469 - val_loss: 5219.5273 - val_r_square: 0.3921\n",
      "Epoch 298/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3790.5825 - r_square: 0.6313 - val_loss: 3921.2471 - val_r_square: 0.5433\n",
      "Epoch 299/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3410.7915 - r_square: 0.6683 - val_loss: 3703.0273 - val_r_square: 0.5687\n",
      "Epoch 300/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3375.6851 - r_square: 0.6717 - val_loss: 4580.9307 - val_r_square: 0.4665\n",
      "Epoch 301/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3415.0825 - r_square: 0.6679 - val_loss: 3752.6248 - val_r_square: 0.5630\n",
      "Epoch 302/500\n",
      "3/3 [==============================] - 0s 69ms/step - loss: 3442.1838 - r_square: 0.6652 - val_loss: 3895.4834 - val_r_square: 0.5463\n",
      "Epoch 303/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3352.8218 - r_square: 0.6739 - val_loss: 4185.8145 - val_r_square: 0.5125\n",
      "Epoch 304/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3364.5518 - r_square: 0.6728 - val_loss: 3604.2180 - val_r_square: 0.5802\n",
      "Epoch 305/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3263.8481 - r_square: 0.6826 - val_loss: 3614.9089 - val_r_square: 0.5790\n",
      "Epoch 306/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3303.7959 - r_square: 0.6787 - val_loss: 4072.0750 - val_r_square: 0.5258\n",
      "Epoch 307/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3211.4744 - r_square: 0.6877 - val_loss: 3535.0605 - val_r_square: 0.5883\n",
      "Epoch 308/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3238.3630 - r_square: 0.6851 - val_loss: 3614.5115 - val_r_square: 0.5790\n",
      "Epoch 309/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3219.9761 - r_square: 0.6868 - val_loss: 4218.3926 - val_r_square: 0.5087\n",
      "Epoch 310/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3317.1497 - r_square: 0.6774 - val_loss: 3967.2871 - val_r_square: 0.5380\n",
      "Epoch 311/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3239.2922 - r_square: 0.6850 - val_loss: 3481.9976 - val_r_square: 0.5945\n",
      "Epoch 312/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3192.0896 - r_square: 0.6896 - val_loss: 3841.2192 - val_r_square: 0.5526\n",
      "Epoch 313/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 3197.4771 - r_square: 0.6890 - val_loss: 4037.8477 - val_r_square: 0.5297\n",
      "Epoch 314/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3195.6318 - r_square: 0.6892 - val_loss: 3530.7246 - val_r_square: 0.5888\n",
      "Epoch 315/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 3150.7385 - r_square: 0.6936 - val_loss: 3804.0547 - val_r_square: 0.5570\n",
      "Epoch 316/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3160.5271 - r_square: 0.6926 - val_loss: 3798.2173 - val_r_square: 0.5577\n",
      "Epoch 317/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3162.8628 - r_square: 0.6924 - val_loss: 3904.7170 - val_r_square: 0.5452\n",
      "Epoch 318/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3173.8594 - r_square: 0.6913 - val_loss: 3572.2693 - val_r_square: 0.5840\n",
      "Epoch 319/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3177.4705 - r_square: 0.6910 - val_loss: 3658.7908 - val_r_square: 0.5739\n",
      "Epoch 320/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3157.8689 - r_square: 0.6929 - val_loss: 3552.1824 - val_r_square: 0.5863\n",
      "Epoch 321/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3163.2261 - r_square: 0.6924 - val_loss: 3799.1699 - val_r_square: 0.5575\n",
      "Epoch 322/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3165.3354 - r_square: 0.6922 - val_loss: 3793.0747 - val_r_square: 0.5583\n",
      "Epoch 323/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3189.9624 - r_square: 0.6898 - val_loss: 3667.2583 - val_r_square: 0.5729\n",
      "Epoch 324/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3157.1965 - r_square: 0.6929 - val_loss: 3537.9792 - val_r_square: 0.5880\n",
      "Epoch 325/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3158.8813 - r_square: 0.6928 - val_loss: 3567.7305 - val_r_square: 0.5845\n",
      "Epoch 326/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3305.1545 - r_square: 0.6786 - val_loss: 3738.1782 - val_r_square: 0.5646\n",
      "Epoch 327/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3277.3347 - r_square: 0.6813 - val_loss: 4210.1763 - val_r_square: 0.5097\n",
      "Epoch 328/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3232.4995 - r_square: 0.6856 - val_loss: 3784.0020 - val_r_square: 0.5593\n",
      "Epoch 329/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3211.7444 - r_square: 0.6876 - val_loss: 3603.5061 - val_r_square: 0.5803\n",
      "Epoch 330/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3113.5083 - r_square: 0.6972 - val_loss: 3784.5940 - val_r_square: 0.5592\n",
      "Epoch 331/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 3117.7227 - r_square: 0.6968 - val_loss: 3720.4321 - val_r_square: 0.5667\n",
      "Epoch 332/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 3104.6260 - r_square: 0.6981 - val_loss: 3535.4021 - val_r_square: 0.5883\n",
      "Epoch 333/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3107.5249 - r_square: 0.6978 - val_loss: 4424.8687 - val_r_square: 0.4847\n",
      "Epoch 334/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3297.3879 - r_square: 0.6793 - val_loss: 4065.2712 - val_r_square: 0.5265\n",
      "Epoch 335/500\n",
      "3/3 [==============================] - 1s 348ms/step - loss: 3230.3096 - r_square: 0.6858 - val_loss: 3766.9602 - val_r_square: 0.5613\n",
      "Epoch 336/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3547.0222 - r_square: 0.6550 - val_loss: 3739.9705 - val_r_square: 0.5644\n",
      "Epoch 337/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3205.9231 - r_square: 0.6882 - val_loss: 4318.4829 - val_r_square: 0.4971\n",
      "Epoch 338/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 3236.8975 - r_square: 0.6852 - val_loss: 3487.9749 - val_r_square: 0.5938\n",
      "Epoch 339/500\n",
      "3/3 [==============================] - 0s 85ms/step - loss: 3330.2883 - r_square: 0.6761 - val_loss: 3617.3794 - val_r_square: 0.5787\n",
      "Epoch 340/500\n",
      "3/3 [==============================] - 0s 82ms/step - loss: 3216.8274 - r_square: 0.6871 - val_loss: 4689.6001 - val_r_square: 0.4538\n",
      "Epoch 341/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3365.5310 - r_square: 0.6727 - val_loss: 3985.1814 - val_r_square: 0.5359\n",
      "Epoch 342/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3206.8708 - r_square: 0.6881 - val_loss: 3522.3220 - val_r_square: 0.5898\n",
      "Epoch 343/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3217.4395 - r_square: 0.6871 - val_loss: 3748.7324 - val_r_square: 0.5634\n",
      "Epoch 344/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3167.1865 - r_square: 0.6920 - val_loss: 3724.2070 - val_r_square: 0.5663\n",
      "Epoch 345/500\n",
      "3/3 [==============================] - 0s 81ms/step - loss: 3127.6445 - r_square: 0.6958 - val_loss: 4179.4648 - val_r_square: 0.5133\n",
      "Epoch 346/500\n",
      "3/3 [==============================] - 0s 83ms/step - loss: 3166.5808 - r_square: 0.6920 - val_loss: 4091.1128 - val_r_square: 0.5235\n",
      "Epoch 347/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 3149.4397 - r_square: 0.6937 - val_loss: 4016.3127 - val_r_square: 0.5323\n",
      "Epoch 348/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3120.4355 - r_square: 0.6965 - val_loss: 3692.0139 - val_r_square: 0.5700\n",
      "Epoch 349/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3140.4282 - r_square: 0.6946 - val_loss: 3774.5972 - val_r_square: 0.5604\n",
      "Epoch 350/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3064.4209 - r_square: 0.7020 - val_loss: 3659.3145 - val_r_square: 0.5738\n",
      "Epoch 351/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3049.9019 - r_square: 0.7034 - val_loss: 3803.0322 - val_r_square: 0.5571\n",
      "Epoch 352/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3269.4490 - r_square: 0.6820 - val_loss: 3883.1899 - val_r_square: 0.5478\n",
      "Epoch 353/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3216.0955 - r_square: 0.6872 - val_loss: 5179.0879 - val_r_square: 0.3968\n",
      "Epoch 354/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3436.8843 - r_square: 0.6657 - val_loss: 3766.3198 - val_r_square: 0.5614\n",
      "Epoch 355/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3190.4187 - r_square: 0.6897 - val_loss: 3553.1450 - val_r_square: 0.5862\n",
      "Epoch 356/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3332.8347 - r_square: 0.6759 - val_loss: 3478.2129 - val_r_square: 0.5949\n",
      "Epoch 357/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3097.8357 - r_square: 0.6987 - val_loss: 4108.3599 - val_r_square: 0.5215\n",
      "Epoch 358/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 3160.4426 - r_square: 0.6926 - val_loss: 4527.9517 - val_r_square: 0.4727\n",
      "Epoch 359/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3448.3450 - r_square: 0.6646 - val_loss: 4059.8562 - val_r_square: 0.5272\n",
      "Epoch 360/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3245.1707 - r_square: 0.6844 - val_loss: 3509.8572 - val_r_square: 0.5912\n",
      "Epoch 361/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3194.5623 - r_square: 0.6893 - val_loss: 3778.4309 - val_r_square: 0.5600\n",
      "Epoch 362/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3058.7446 - r_square: 0.7025 - val_loss: 4075.7751 - val_r_square: 0.5253\n",
      "Epoch 363/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3051.0906 - r_square: 0.7033 - val_loss: 3582.0483 - val_r_square: 0.5828\n",
      "Epoch 364/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3167.1125 - r_square: 0.6920 - val_loss: 3681.7415 - val_r_square: 0.5712\n",
      "Epoch 365/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3339.0786 - r_square: 0.6753 - val_loss: 4067.3821 - val_r_square: 0.5263\n",
      "Epoch 366/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3281.0579 - r_square: 0.6809 - val_loss: 4367.4819 - val_r_square: 0.4914\n",
      "Epoch 367/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3107.8496 - r_square: 0.6977 - val_loss: 3500.7734 - val_r_square: 0.5923\n",
      "Epoch 368/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3080.1287 - r_square: 0.7004 - val_loss: 3733.3352 - val_r_square: 0.5652\n",
      "Epoch 369/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3093.2263 - r_square: 0.6992 - val_loss: 3960.3657 - val_r_square: 0.5388\n",
      "Epoch 370/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3056.0818 - r_square: 0.7028 - val_loss: 3801.4810 - val_r_square: 0.5573\n",
      "Epoch 371/500\n",
      "3/3 [==============================] - 0s 55ms/step - loss: 2986.6741 - r_square: 0.7095 - val_loss: 3567.1138 - val_r_square: 0.5846\n",
      "Epoch 372/500\n",
      "3/3 [==============================] - 0s 55ms/step - loss: 3076.6458 - r_square: 0.7008 - val_loss: 3618.7148 - val_r_square: 0.5786\n",
      "Epoch 373/500\n",
      "3/3 [==============================] - 0s 62ms/step - loss: 3017.4746 - r_square: 0.7065 - val_loss: 3569.5840 - val_r_square: 0.5843\n",
      "Epoch 374/500\n",
      "3/3 [==============================] - 0s 55ms/step - loss: 3026.3430 - r_square: 0.7057 - val_loss: 3534.5881 - val_r_square: 0.5884\n",
      "Epoch 375/500\n",
      "3/3 [==============================] - 0s 61ms/step - loss: 3017.0952 - r_square: 0.7066 - val_loss: 4151.0728 - val_r_square: 0.5166\n",
      "Epoch 376/500\n",
      "3/3 [==============================] - 0s 59ms/step - loss: 3190.8259 - r_square: 0.6897 - val_loss: 4558.4580 - val_r_square: 0.4691\n",
      "Epoch 377/500\n",
      "3/3 [==============================] - 0s 82ms/step - loss: 3374.3462 - r_square: 0.6718 - val_loss: 4002.3250 - val_r_square: 0.5339\n",
      "Epoch 378/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3282.8689 - r_square: 0.6807 - val_loss: 3708.5457 - val_r_square: 0.5681\n",
      "Epoch 379/500\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 3328.6523 - r_square: 0.6763 - val_loss: 3811.5208 - val_r_square: 0.5561\n",
      "Epoch 380/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3067.5635 - r_square: 0.7017 - val_loss: 4703.1567 - val_r_square: 0.4523\n",
      "Epoch 381/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3281.4514 - r_square: 0.6809 - val_loss: 3867.7175 - val_r_square: 0.5496\n",
      "Epoch 382/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3080.4746 - r_square: 0.7004 - val_loss: 3809.0740 - val_r_square: 0.5564\n",
      "Epoch 383/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3094.0884 - r_square: 0.6991 - val_loss: 3789.4365 - val_r_square: 0.5587\n",
      "Epoch 384/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3052.9070 - r_square: 0.7031 - val_loss: 3660.2058 - val_r_square: 0.5737\n",
      "Epoch 385/500\n",
      "3/3 [==============================] - 0s 66ms/step - loss: 2976.3359 - r_square: 0.7105 - val_loss: 4032.2620 - val_r_square: 0.5304\n",
      "Epoch 386/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3024.6230 - r_square: 0.7058 - val_loss: 3590.9773 - val_r_square: 0.5818\n",
      "Epoch 387/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3038.8433 - r_square: 0.7045 - val_loss: 3606.2568 - val_r_square: 0.5800\n",
      "Epoch 388/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3045.6943 - r_square: 0.7038 - val_loss: 3714.6780 - val_r_square: 0.5674\n",
      "Epoch 389/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 2977.7615 - r_square: 0.7104 - val_loss: 3623.8984 - val_r_square: 0.5780\n",
      "Epoch 390/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 2967.2063 - r_square: 0.7114 - val_loss: 3641.5725 - val_r_square: 0.5759\n",
      "Epoch 391/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 2957.8613 - r_square: 0.7123 - val_loss: 4310.4233 - val_r_square: 0.4980\n",
      "Epoch 392/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3169.7920 - r_square: 0.6917 - val_loss: 5012.4570 - val_r_square: 0.4162\n",
      "Epoch 393/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3337.0144 - r_square: 0.6755 - val_loss: 3451.0239 - val_r_square: 0.5981\n",
      "Epoch 394/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3503.1311 - r_square: 0.6593 - val_loss: 3712.5933 - val_r_square: 0.5676\n",
      "Epoch 395/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3357.2588 - r_square: 0.6735 - val_loss: 3738.1558 - val_r_square: 0.5646\n",
      "Epoch 396/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3328.3003 - r_square: 0.6763 - val_loss: 5025.9902 - val_r_square: 0.4147\n",
      "Epoch 397/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3334.1997 - r_square: 0.6757 - val_loss: 3970.3374 - val_r_square: 0.5376\n",
      "Epoch 398/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3303.5137 - r_square: 0.6787 - val_loss: 4000.9465 - val_r_square: 0.5340\n",
      "Epoch 399/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3649.8022 - r_square: 0.6450 - val_loss: 3584.4937 - val_r_square: 0.5825\n",
      "Epoch 400/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3111.6084 - r_square: 0.6974 - val_loss: 4692.8350 - val_r_square: 0.4535\n",
      "Epoch 401/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 3151.3367 - r_square: 0.6935 - val_loss: 3708.3418 - val_r_square: 0.5681\n",
      "Epoch 402/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3229.5225 - r_square: 0.6859 - val_loss: 3622.8318 - val_r_square: 0.5781\n",
      "Epoch 403/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3216.7598 - r_square: 0.6872 - val_loss: 3931.7849 - val_r_square: 0.5421\n",
      "Epoch 404/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3111.0322 - r_square: 0.6974 - val_loss: 5101.4478 - val_r_square: 0.4059\n",
      "Epoch 405/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3534.3635 - r_square: 0.6563 - val_loss: 3740.5068 - val_r_square: 0.5644\n",
      "Epoch 406/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3222.7520 - r_square: 0.6866 - val_loss: 3635.2776 - val_r_square: 0.5766\n",
      "Epoch 407/500\n",
      "3/3 [==============================] - 0s 82ms/step - loss: 3354.8926 - r_square: 0.6737 - val_loss: 3788.6462 - val_r_square: 0.5588\n",
      "Epoch 408/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 3083.2576 - r_square: 0.7001 - val_loss: 3504.6924 - val_r_square: 0.5918\n",
      "Epoch 409/500\n",
      "3/3 [==============================] - 0s 68ms/step - loss: 3039.6040 - r_square: 0.7044 - val_loss: 3901.0913 - val_r_square: 0.5457\n",
      "Epoch 410/500\n",
      "3/3 [==============================] - 0s 100ms/step - loss: 3004.0068 - r_square: 0.7078 - val_loss: 4009.8103 - val_r_square: 0.5330\n",
      "Epoch 411/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 2947.5449 - r_square: 0.7133 - val_loss: 3684.1921 - val_r_square: 0.5709\n",
      "Epoch 412/500\n",
      "3/3 [==============================] - 0s 83ms/step - loss: 3069.4468 - r_square: 0.7015 - val_loss: 3722.7317 - val_r_square: 0.5664\n",
      "Epoch 413/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 3155.5774 - r_square: 0.6931 - val_loss: 3608.8169 - val_r_square: 0.5797\n",
      "Epoch 414/500\n",
      "3/3 [==============================] - 0s 85ms/step - loss: 3001.7368 - r_square: 0.7081 - val_loss: 4026.4258 - val_r_square: 0.5311\n",
      "Epoch 415/500\n",
      "3/3 [==============================] - 0s 81ms/step - loss: 3002.8833 - r_square: 0.7080 - val_loss: 3796.4536 - val_r_square: 0.5579\n",
      "Epoch 416/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3043.0867 - r_square: 0.7040 - val_loss: 3623.7690 - val_r_square: 0.5780\n",
      "Epoch 417/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3006.4214 - r_square: 0.7076 - val_loss: 3842.1265 - val_r_square: 0.5525\n",
      "Epoch 418/500\n",
      "3/3 [==============================] - 0s 81ms/step - loss: 3250.8145 - r_square: 0.6838 - val_loss: 3796.8940 - val_r_square: 0.5578\n",
      "Epoch 419/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3130.7585 - r_square: 0.6955 - val_loss: 4500.3008 - val_r_square: 0.4759\n",
      "Epoch 420/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3263.4592 - r_square: 0.6826 - val_loss: 5240.7563 - val_r_square: 0.3897\n",
      "Epoch 421/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 3546.7229 - r_square: 0.6551 - val_loss: 3669.0193 - val_r_square: 0.5727\n",
      "Epoch 422/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3372.4014 - r_square: 0.6720 - val_loss: 3878.6489 - val_r_square: 0.5483\n",
      "Epoch 423/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 3426.1064 - r_square: 0.6668 - val_loss: 4957.0273 - val_r_square: 0.4227\n",
      "Epoch 424/500\n",
      "3/3 [==============================] - 1s 288ms/step - loss: 3132.0745 - r_square: 0.6954 - val_loss: 3717.6384 - val_r_square: 0.5670\n",
      "Epoch 425/500\n",
      "3/3 [==============================] - 0s 136ms/step - loss: 3058.4858 - r_square: 0.7025 - val_loss: 3588.9792 - val_r_square: 0.5820\n",
      "Epoch 426/500\n",
      "3/3 [==============================] - 0s 88ms/step - loss: 3096.0840 - r_square: 0.6989 - val_loss: 3824.1528 - val_r_square: 0.5546\n",
      "Epoch 427/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 2975.7385 - r_square: 0.7106 - val_loss: 4138.4883 - val_r_square: 0.5180\n",
      "Epoch 428/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 3025.3025 - r_square: 0.7058 - val_loss: 4120.8350 - val_r_square: 0.5201\n",
      "Epoch 429/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3062.6248 - r_square: 0.7021 - val_loss: 3499.3916 - val_r_square: 0.5925\n",
      "Epoch 430/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 2971.7085 - r_square: 0.7110 - val_loss: 3769.1880 - val_r_square: 0.5610\n",
      "Epoch 431/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3023.5178 - r_square: 0.7059 - val_loss: 4291.8760 - val_r_square: 0.5002\n",
      "Epoch 432/500\n",
      "3/3 [==============================] - 0s 87ms/step - loss: 2983.6245 - r_square: 0.7098 - val_loss: 3825.5923 - val_r_square: 0.5545\n",
      "Epoch 433/500\n",
      "3/3 [==============================] - 0s 84ms/step - loss: 2956.5500 - r_square: 0.7125 - val_loss: 3790.0757 - val_r_square: 0.5586\n",
      "Epoch 434/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 2872.0662 - r_square: 0.7207 - val_loss: 3675.5767 - val_r_square: 0.5719\n",
      "Epoch 435/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 2973.2075 - r_square: 0.7108 - val_loss: 3807.5066 - val_r_square: 0.5566\n",
      "Epoch 436/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 2945.8052 - r_square: 0.7135 - val_loss: 3689.3152 - val_r_square: 0.5703\n",
      "Epoch 437/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 2920.8679 - r_square: 0.7159 - val_loss: 3764.3523 - val_r_square: 0.5616\n",
      "Epoch 438/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 2939.8018 - r_square: 0.7141 - val_loss: 3627.0347 - val_r_square: 0.5776\n",
      "Epoch 439/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 2866.7292 - r_square: 0.7212 - val_loss: 4965.6162 - val_r_square: 0.4217\n",
      "Epoch 440/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3264.1086 - r_square: 0.6825 - val_loss: 4085.9885 - val_r_square: 0.5241\n",
      "Epoch 441/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 3147.1025 - r_square: 0.6939 - val_loss: 3601.6743 - val_r_square: 0.5805\n",
      "Epoch 442/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 3223.3042 - r_square: 0.6865 - val_loss: 3683.0803 - val_r_square: 0.5711\n",
      "Epoch 443/500\n",
      "3/3 [==============================] - 0s 89ms/step - loss: 3073.0801 - r_square: 0.7011 - val_loss: 3917.5828 - val_r_square: 0.5437\n",
      "Epoch 444/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 2929.1917 - r_square: 0.7151 - val_loss: 3907.4746 - val_r_square: 0.5449\n",
      "Epoch 445/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 2864.6189 - r_square: 0.7214 - val_loss: 3747.0063 - val_r_square: 0.5636\n",
      "Epoch 446/500\n",
      "3/3 [==============================] - 0s 83ms/step - loss: 3070.2903 - r_square: 0.7014 - val_loss: 3573.4568 - val_r_square: 0.5838\n",
      "Epoch 447/500\n",
      "3/3 [==============================] - 0s 84ms/step - loss: 2998.3411 - r_square: 0.7084 - val_loss: 4224.3076 - val_r_square: 0.5080\n",
      "Epoch 448/500\n",
      "3/3 [==============================] - 0s 82ms/step - loss: 2997.1418 - r_square: 0.7085 - val_loss: 4408.7793 - val_r_square: 0.4865\n",
      "Epoch 449/500\n",
      "3/3 [==============================] - 0s 81ms/step - loss: 3043.0308 - r_square: 0.7040 - val_loss: 4021.8362 - val_r_square: 0.5316\n",
      "Epoch 450/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 2966.8125 - r_square: 0.7115 - val_loss: 3977.9089 - val_r_square: 0.5367\n",
      "Epoch 451/500\n",
      "3/3 [==============================] - 0s 85ms/step - loss: 2853.3201 - r_square: 0.7225 - val_loss: 3997.6589 - val_r_square: 0.5344\n",
      "Epoch 452/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 2883.9783 - r_square: 0.7195 - val_loss: 3727.2883 - val_r_square: 0.5659\n",
      "Epoch 453/500\n",
      "3/3 [==============================] - 0s 86ms/step - loss: 2835.1672 - r_square: 0.7243 - val_loss: 3739.5020 - val_r_square: 0.5645\n",
      "Epoch 454/500\n",
      "3/3 [==============================] - 0s 79ms/step - loss: 2811.8245 - r_square: 0.7265 - val_loss: 3766.0078 - val_r_square: 0.5614\n",
      "Epoch 455/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 2925.2261 - r_square: 0.7155 - val_loss: 4737.9111 - val_r_square: 0.4482\n",
      "Epoch 456/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3003.5881 - r_square: 0.7079 - val_loss: 3566.8967 - val_r_square: 0.5846\n",
      "Epoch 457/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 2878.7837 - r_square: 0.7200 - val_loss: 3691.8440 - val_r_square: 0.5700\n",
      "Epoch 458/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 2806.4141 - r_square: 0.7271 - val_loss: 3933.5688 - val_r_square: 0.5419\n",
      "Epoch 459/500\n",
      "3/3 [==============================] - 0s 84ms/step - loss: 2811.0586 - r_square: 0.7266 - val_loss: 3758.6899 - val_r_square: 0.5623\n",
      "Epoch 460/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 2806.6130 - r_square: 0.7270 - val_loss: 4004.6086 - val_r_square: 0.5336\n",
      "Epoch 461/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 2865.6577 - r_square: 0.7213 - val_loss: 3693.8882 - val_r_square: 0.5698\n",
      "Epoch 462/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 2938.7900 - r_square: 0.7142 - val_loss: 3604.4675 - val_r_square: 0.5802\n",
      "Epoch 463/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 3007.5239 - r_square: 0.7075 - val_loss: 4057.8604 - val_r_square: 0.5274\n",
      "Epoch 464/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 2937.9165 - r_square: 0.7143 - val_loss: 4171.7134 - val_r_square: 0.5142\n",
      "Epoch 465/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 2872.0415 - r_square: 0.7207 - val_loss: 3749.8210 - val_r_square: 0.5633\n",
      "Epoch 466/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 2870.1865 - r_square: 0.7209 - val_loss: 3541.3062 - val_r_square: 0.5876\n",
      "Epoch 467/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 2852.1206 - r_square: 0.7226 - val_loss: 3780.1558 - val_r_square: 0.5598\n",
      "Epoch 468/500\n",
      "3/3 [==============================] - 0s 71ms/step - loss: 2781.6621 - r_square: 0.7295 - val_loss: 3813.9646 - val_r_square: 0.5558\n",
      "Epoch 469/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 2838.1348 - r_square: 0.7240 - val_loss: 4303.7788 - val_r_square: 0.4988\n",
      "Epoch 470/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 2955.6975 - r_square: 0.7125 - val_loss: 4471.0049 - val_r_square: 0.4793\n",
      "Epoch 471/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 3072.0344 - r_square: 0.7012 - val_loss: 3985.5178 - val_r_square: 0.5358\n",
      "Epoch 472/500\n",
      "3/3 [==============================] - 0s 70ms/step - loss: 3149.5269 - r_square: 0.6937 - val_loss: 3799.1577 - val_r_square: 0.5575\n",
      "Epoch 473/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 2851.9766 - r_square: 0.7226 - val_loss: 3542.6274 - val_r_square: 0.5874\n",
      "Epoch 474/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 2911.0508 - r_square: 0.7169 - val_loss: 3553.4441 - val_r_square: 0.5862\n",
      "Epoch 475/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 2928.8396 - r_square: 0.7152 - val_loss: 4209.9248 - val_r_square: 0.5097\n",
      "Epoch 476/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 2861.2888 - r_square: 0.7217 - val_loss: 4090.8435 - val_r_square: 0.5236\n",
      "Epoch 477/500\n",
      "3/3 [==============================] - 0s 72ms/step - loss: 2922.4788 - r_square: 0.7158 - val_loss: 3895.6343 - val_r_square: 0.5463\n",
      "Epoch 478/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 2840.0420 - r_square: 0.7238 - val_loss: 3843.8318 - val_r_square: 0.5523\n",
      "Epoch 479/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 3045.5569 - r_square: 0.7038 - val_loss: 3683.8535 - val_r_square: 0.5710\n",
      "Epoch 480/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3098.3611 - r_square: 0.6987 - val_loss: 3987.5366 - val_r_square: 0.5356\n",
      "Epoch 481/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 2983.6675 - r_square: 0.7098 - val_loss: 4756.7905 - val_r_square: 0.4460\n",
      "Epoch 482/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 3024.1174 - r_square: 0.7059 - val_loss: 4418.4502 - val_r_square: 0.4854\n",
      "Epoch 483/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 2887.9241 - r_square: 0.7191 - val_loss: 3542.5134 - val_r_square: 0.5874\n",
      "Epoch 484/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 2875.8906 - r_square: 0.7203 - val_loss: 3554.3098 - val_r_square: 0.5861\n",
      "Epoch 485/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 2907.4749 - r_square: 0.7172 - val_loss: 3946.1836 - val_r_square: 0.5404\n",
      "Epoch 486/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 2815.8958 - r_square: 0.7261 - val_loss: 4279.7090 - val_r_square: 0.5016\n",
      "Epoch 487/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 2748.8455 - r_square: 0.7327 - val_loss: 3642.2244 - val_r_square: 0.5758\n",
      "Epoch 488/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 2806.5769 - r_square: 0.7270 - val_loss: 3772.8479 - val_r_square: 0.5606\n",
      "Epoch 489/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 2737.9370 - r_square: 0.7337 - val_loss: 3771.7747 - val_r_square: 0.5607\n",
      "Epoch 490/500\n",
      "3/3 [==============================] - 0s 86ms/step - loss: 2805.9214 - r_square: 0.7271 - val_loss: 3516.2144 - val_r_square: 0.5905\n",
      "Epoch 491/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 2787.5195 - r_square: 0.7289 - val_loss: 3565.8613 - val_r_square: 0.5847\n",
      "Epoch 492/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 2910.5596 - r_square: 0.7169 - val_loss: 3638.6162 - val_r_square: 0.5762\n",
      "Epoch 493/500\n",
      "3/3 [==============================] - 0s 76ms/step - loss: 2778.6140 - r_square: 0.7298 - val_loss: 3788.6663 - val_r_square: 0.5588\n",
      "Epoch 494/500\n",
      "3/3 [==============================] - 0s 80ms/step - loss: 2689.5027 - r_square: 0.7384 - val_loss: 4055.9441 - val_r_square: 0.5276\n",
      "Epoch 495/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 2712.5503 - r_square: 0.7362 - val_loss: 3697.6304 - val_r_square: 0.5694\n",
      "Epoch 496/500\n",
      "3/3 [==============================] - 0s 73ms/step - loss: 2670.6196 - r_square: 0.7403 - val_loss: 3588.8335 - val_r_square: 0.5820\n",
      "Epoch 497/500\n",
      "3/3 [==============================] - 0s 74ms/step - loss: 2744.8027 - r_square: 0.7331 - val_loss: 4200.4653 - val_r_square: 0.5108\n",
      "Epoch 498/500\n",
      "3/3 [==============================] - 0s 75ms/step - loss: 2807.9065 - r_square: 0.7269 - val_loss: 4643.3208 - val_r_square: 0.4592\n",
      "Epoch 499/500\n",
      "3/3 [==============================] - 0s 78ms/step - loss: 2923.5996 - r_square: 0.7157 - val_loss: 4521.7788 - val_r_square: 0.4734\n",
      "Epoch 500/500\n",
      "3/3 [==============================] - 0s 77ms/step - loss: 2960.8330 - r_square: 0.7120 - val_loss: 3793.9604 - val_r_square: 0.5581\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.layers import LSTM, Dense, Bidirectional\n",
    "\n",
    "model_bd = Sequential()\n",
    "model_bd.add(Bidirectional(LSTM(64, activation='relu', return_sequences=True), input_shape=(6,5)))\n",
    "model_bd.add(Bidirectional(LSTM(64, activation='relu')))\n",
    "model_bd.add(Dense(1, activation='linear'))\n",
    "\n",
    "model_bd.compile(loss='mean_squared_error', optimizer=tf.keras.optimizers.Adam(learning_rate=0.01), metrics=[tfa.metrics.RSquare()])\n",
    "\n",
    "\n",
    "history_bd = model_bd.fit(X_train_rs, y_train1, epochs=500, batch_size=350, validation_data=(X_val_rs, y_val1))\n",
    "\n",
    "#history_bd = model_bd.fit(X_train_rs, y_train1, epochs=500, batch_size=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 312,
   "id": "b85382ec-354e-48ce-8eae-804b15e451cc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 0s 5ms/step\n",
      "51.76553676949517\n"
     ]
    }
   ],
   "source": [
    "y_pred_bd_tr = model_bd.predict(X_train_rs)\n",
    "bd_rmse_train = np.sqrt(mean_squared_error(y_train1, y_pred_bd_tr))\n",
    "print(bd_rmse_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 313,
   "id": "0ef4ba4c-6144-42c9-b46b-5ba7de36d373",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3/3 [==============================] - 0s 2ms/step - loss: 3981.6562 - r_square: 0.5797\n"
     ]
    }
   ],
   "source": [
    "results_bd = model_bd.evaluate(X_test_rs, y_test1, batch_size=128)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 314,
   "id": "83996940-0858-44c6-bea9-6ec08f65f930",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9/9 [==============================] - 0s 7ms/step\n",
      "63.100363603384885\n"
     ]
    }
   ],
   "source": [
    "y_pred_bd = model_bd.predict(X_test_rs)\n",
    "bd_rmse = np.sqrt(mean_squared_error(y_test1, y_pred_bd))\n",
    "print(bd_rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "c6614f4a-0b75-4f32-80e7-d051ec40547e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history_bd)\n",
    "plt.savefig('PlotLoss_BD.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "id": "c2a5aeef-c4e7-4911-9b1b-6b1a85a5a29e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "error = y_pred_bd - y_test1\n",
    "plt.hist(error, bins=25)\n",
    "plt.xlabel('Prediction Error')\n",
    "_ = plt.ylabel('Count')\n",
    "\n",
    "fig = matplotlib.pyplot.gcf()\n",
    "fig.set_size_inches(10,10)\n",
    "fig.savefig('pred_error_bd.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ba154f79-b372-4ea0-8a53-5f1680c3b230",
   "metadata": {},
   "source": [
    "## ------------- Bidirectional LSTM with Dropout ---------------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 381,
   "id": "a1a88010-ebdf-4aea-b91d-bb30fa63623c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "3/3 [==============================] - 9s 538ms/step - loss: 390087.0625 - r_square: -36.9377 - val_loss: 137962.4062 - val_r_square: -15.0670\n",
      "Epoch 2/500\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 14451583.0000 - r_square: -1404.4949 - val_loss: 396591.1562 - val_r_square: -45.1874\n",
      "Epoch 3/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 379205.0938 - r_square: -35.8792 - val_loss: 423854.6250 - val_r_square: -48.3626\n",
      "Epoch 4/500\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 391966.0312 - r_square: -37.1203 - val_loss: 424813.5625 - val_r_square: -48.4743\n",
      "Epoch 5/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 392592.2812 - r_square: -37.1811 - val_loss: 425020.1562 - val_r_square: -48.4983\n",
      "Epoch 6/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 392743.6875 - r_square: -37.1966 - val_loss: 425078.6250 - val_r_square: -48.5051\n",
      "Epoch 7/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 392787.8125 - r_square: -37.2005 - val_loss: 425089.9688 - val_r_square: -48.5065\n",
      "Epoch 8/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 392794.3438 - r_square: -37.2008 - val_loss: 425081.2812 - val_r_square: -48.5055\n",
      "Epoch 9/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 392784.5312 - r_square: -37.2001 - val_loss: 425063.2188 - val_r_square: -48.5034\n",
      "Epoch 10/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 392767.2188 - r_square: -37.1984 - val_loss: 425040.8125 - val_r_square: -48.5007\n",
      "Epoch 11/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 392745.8438 - r_square: -37.1961 - val_loss: 425016.2188 - val_r_square: -48.4979\n",
      "Epoch 12/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 392722.5938 - r_square: -37.1939 - val_loss: 424990.3438 - val_r_square: -48.4949\n",
      "Epoch 13/500\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 392697.8438 - r_square: -37.1914 - val_loss: 424963.4375 - val_r_square: -48.4917\n",
      "Epoch 14/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 392672.0312 - r_square: -37.1893 - val_loss: 424935.7812 - val_r_square: -48.4885\n",
      "Epoch 15/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 392645.5000 - r_square: -37.1862 - val_loss: 424907.5312 - val_r_square: -48.4852\n",
      "Epoch 16/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 392618.2812 - r_square: -37.1839 - val_loss: 424878.6875 - val_r_square: -48.4819\n",
      "Epoch 17/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 392590.5938 - r_square: -37.1808 - val_loss: 424849.2812 - val_r_square: -48.4785\n",
      "Epoch 18/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 392562.3438 - r_square: -37.1784 - val_loss: 424819.5938 - val_r_square: -48.4750\n",
      "Epoch 19/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 392533.8125 - r_square: -37.1756 - val_loss: 424789.4062 - val_r_square: -48.4715\n",
      "Epoch 20/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 392504.7812 - r_square: -37.1730 - val_loss: 424758.9375 - val_r_square: -48.4679\n",
      "Epoch 21/500\n",
      "3/3 [==============================] - 0s 104ms/step - loss: 392475.4688 - r_square: -37.1702 - val_loss: 424728.1250 - val_r_square: -48.4644\n",
      "Epoch 22/500\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 392445.9062 - r_square: -37.1670 - val_loss: 424697.0625 - val_r_square: -48.4607\n",
      "Epoch 23/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 392416.0312 - r_square: -37.1640 - val_loss: 424665.6875 - val_r_square: -48.4571\n",
      "Epoch 24/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 392385.8750 - r_square: -37.1615 - val_loss: 424634.0625 - val_r_square: -48.4534\n",
      "Epoch 25/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 392355.6250 - r_square: -37.1585 - val_loss: 424602.2500 - val_r_square: -48.4497\n",
      "Epoch 26/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 392324.9688 - r_square: -37.1553 - val_loss: 424570.2500 - val_r_square: -48.4460\n",
      "Epoch 27/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 392294.2188 - r_square: -37.1524 - val_loss: 424537.9062 - val_r_square: -48.4422\n",
      "Epoch 28/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 392263.2812 - r_square: -37.1494 - val_loss: 424505.5312 - val_r_square: -48.4384\n",
      "Epoch 29/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 392232.2188 - r_square: -37.1464 - val_loss: 424472.9062 - val_r_square: -48.4346\n",
      "Epoch 30/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 392200.9375 - r_square: -37.1435 - val_loss: 424440.1250 - val_r_square: -48.4308\n",
      "Epoch 31/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 392169.4688 - r_square: -37.1404 - val_loss: 424407.2188 - val_r_square: -48.4270\n",
      "Epoch 32/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 392137.9062 - r_square: -37.1374 - val_loss: 424374.1562 - val_r_square: -48.4231\n",
      "Epoch 33/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 392106.1250 - r_square: -37.1343 - val_loss: 424340.9062 - val_r_square: -48.4193\n",
      "Epoch 34/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 392074.3125 - r_square: -37.1310 - val_loss: 424307.5938 - val_r_square: -48.4154\n",
      "Epoch 35/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 392042.2500 - r_square: -37.1277 - val_loss: 424274.1250 - val_r_square: -48.4115\n",
      "Epoch 36/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 392010.1875 - r_square: -37.1248 - val_loss: 424240.5625 - val_r_square: -48.4076\n",
      "Epoch 37/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 391978.0000 - r_square: -37.1218 - val_loss: 424206.8125 - val_r_square: -48.4037\n",
      "Epoch 38/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 391945.5625 - r_square: -37.1184 - val_loss: 424173.0312 - val_r_square: -48.3997\n",
      "Epoch 39/500\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 391913.1562 - r_square: -37.1155 - val_loss: 424139.0625 - val_r_square: -48.3958\n",
      "Epoch 40/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 391880.5938 - r_square: -37.1121 - val_loss: 424105.0000 - val_r_square: -48.3918\n",
      "Epoch 41/500\n",
      "3/3 [==============================] - 0s 164ms/step - loss: 391848.0000 - r_square: -37.1092 - val_loss: 424070.8750 - val_r_square: -48.3878\n",
      "Epoch 42/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 391815.2188 - r_square: -37.1060 - val_loss: 424036.6562 - val_r_square: -48.3838\n",
      "Epoch 43/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 391782.3125 - r_square: -37.1026 - val_loss: 424002.3750 - val_r_square: -48.3798\n",
      "Epoch 44/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 391749.5000 - r_square: -37.0993 - val_loss: 423967.9375 - val_r_square: -48.3758\n",
      "Epoch 45/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 391716.4688 - r_square: -37.0962 - val_loss: 423933.4375 - val_r_square: -48.3718\n",
      "Epoch 46/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 391683.3750 - r_square: -37.0928 - val_loss: 423898.8750 - val_r_square: -48.3678\n",
      "Epoch 47/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 391650.2500 - r_square: -37.0898 - val_loss: 423864.2500 - val_r_square: -48.3638\n",
      "Epoch 48/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 391617.1250 - r_square: -37.0864 - val_loss: 423829.5000 - val_r_square: -48.3597\n",
      "Epoch 49/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 391583.7812 - r_square: -37.0833 - val_loss: 423794.7188 - val_r_square: -48.3557\n",
      "Epoch 50/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 391550.4062 - r_square: -37.0802 - val_loss: 423759.9062 - val_r_square: -48.3516\n",
      "Epoch 51/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 391517.0000 - r_square: -37.0767 - val_loss: 423725.0000 - val_r_square: -48.3475\n",
      "Epoch 52/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 391483.5312 - r_square: -37.0736 - val_loss: 423690.0312 - val_r_square: -48.3435\n",
      "Epoch 53/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 391449.9062 - r_square: -37.0705 - val_loss: 423655.0312 - val_r_square: -48.3394\n",
      "Epoch 54/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 391416.4062 - r_square: -37.0671 - val_loss: 423619.9062 - val_r_square: -48.3353\n",
      "Epoch 55/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 391382.7812 - r_square: -37.0635 - val_loss: 423584.7188 - val_r_square: -48.3312\n",
      "Epoch 56/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 391348.9688 - r_square: -37.0605 - val_loss: 423549.5000 - val_r_square: -48.3271\n",
      "Epoch 57/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 391315.2812 - r_square: -37.0570 - val_loss: 423514.2500 - val_r_square: -48.3230\n",
      "Epoch 58/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 391281.3125 - r_square: -37.0538 - val_loss: 423478.8750 - val_r_square: -48.3189\n",
      "Epoch 59/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 391247.6250 - r_square: -37.0512 - val_loss: 423443.5000 - val_r_square: -48.3148\n",
      "Epoch 60/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 391213.6250 - r_square: -37.0475 - val_loss: 423408.1250 - val_r_square: -48.3106\n",
      "Epoch 61/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 391179.7500 - r_square: -37.0442 - val_loss: 423372.5625 - val_r_square: -48.3065\n",
      "Epoch 62/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 391145.6562 - r_square: -37.0405 - val_loss: 423337.0625 - val_r_square: -48.3024\n",
      "Epoch 63/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 391111.6875 - r_square: -37.0372 - val_loss: 423301.5000 - val_r_square: -48.2982\n",
      "Epoch 64/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 391077.5000 - r_square: -37.0340 - val_loss: 423265.8125 - val_r_square: -48.2941\n",
      "Epoch 65/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 391043.2812 - r_square: -37.0306 - val_loss: 423230.1875 - val_r_square: -48.2899\n",
      "Epoch 66/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 391009.2500 - r_square: -37.0275 - val_loss: 423194.4688 - val_r_square: -48.2858\n",
      "Epoch 67/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 390975.0000 - r_square: -37.0240 - val_loss: 423158.7500 - val_r_square: -48.2816\n",
      "Epoch 68/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 390940.7188 - r_square: -37.0205 - val_loss: 423122.9375 - val_r_square: -48.2774\n",
      "Epoch 69/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 390906.4062 - r_square: -37.0173 - val_loss: 423087.1250 - val_r_square: -48.2733\n",
      "Epoch 70/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 390872.0938 - r_square: -37.0140 - val_loss: 423051.3125 - val_r_square: -48.2691\n",
      "Epoch 71/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 390837.6562 - r_square: -37.0105 - val_loss: 423015.4375 - val_r_square: -48.2649\n",
      "Epoch 72/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 390803.3750 - r_square: -37.0071 - val_loss: 422979.5625 - val_r_square: -48.2607\n",
      "Epoch 73/500\n",
      "3/3 [==============================] - 0s 161ms/step - loss: 390768.9375 - r_square: -37.0038 - val_loss: 422943.5938 - val_r_square: -48.2565\n",
      "Epoch 74/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 390734.5000 - r_square: -37.0007 - val_loss: 422907.5625 - val_r_square: -48.2524\n",
      "Epoch 75/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 390699.9062 - r_square: -36.9973 - val_loss: 422871.5312 - val_r_square: -48.2482\n",
      "Epoch 76/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 390665.3125 - r_square: -36.9938 - val_loss: 422835.5000 - val_r_square: -48.2440\n",
      "Epoch 77/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 390630.7812 - r_square: -36.9905 - val_loss: 422799.4062 - val_r_square: -48.2398\n",
      "Epoch 78/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 390596.3125 - r_square: -36.9873 - val_loss: 422763.2500 - val_r_square: -48.2356\n",
      "Epoch 79/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 390561.5938 - r_square: -36.9838 - val_loss: 422727.0938 - val_r_square: -48.2313\n",
      "Epoch 80/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 390526.9062 - r_square: -36.9806 - val_loss: 422690.9375 - val_r_square: -48.2271\n",
      "Epoch 81/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 390492.3125 - r_square: -36.9772 - val_loss: 422654.7188 - val_r_square: -48.2229\n",
      "Epoch 82/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 390457.5938 - r_square: -36.9738 - val_loss: 422618.5000 - val_r_square: -48.2187\n",
      "Epoch 83/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 390422.7812 - r_square: -36.9702 - val_loss: 422582.2188 - val_r_square: -48.2145\n",
      "Epoch 84/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 390388.0938 - r_square: -36.9669 - val_loss: 422545.9688 - val_r_square: -48.2102\n",
      "Epoch 85/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 390353.2812 - r_square: -36.9638 - val_loss: 422509.5938 - val_r_square: -48.2060\n",
      "Epoch 86/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 390318.5312 - r_square: -36.9604 - val_loss: 422473.2188 - val_r_square: -48.2018\n",
      "Epoch 87/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 390283.6250 - r_square: -36.9568 - val_loss: 422436.9062 - val_r_square: -48.1975\n",
      "Epoch 88/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 390248.8125 - r_square: -36.9536 - val_loss: 422400.5000 - val_r_square: -48.1933\n",
      "Epoch 89/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 390213.8750 - r_square: -36.9500 - val_loss: 422364.0625 - val_r_square: -48.1891\n",
      "Epoch 90/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 390179.0938 - r_square: -36.9465 - val_loss: 422327.6250 - val_r_square: -48.1848\n",
      "Epoch 91/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 390144.2188 - r_square: -36.9431 - val_loss: 422291.1875 - val_r_square: -48.1806\n",
      "Epoch 92/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 390109.1250 - r_square: -36.9397 - val_loss: 422254.6875 - val_r_square: -48.1763\n",
      "Epoch 93/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 390074.1562 - r_square: -36.9363 - val_loss: 422218.2500 - val_r_square: -48.1721\n",
      "Epoch 94/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 390039.2812 - r_square: -36.9329 - val_loss: 422181.7500 - val_r_square: -48.1678\n",
      "Epoch 95/500\n",
      "3/3 [==============================] - 0s 161ms/step - loss: 390004.2500 - r_square: -36.9293 - val_loss: 422145.1562 - val_r_square: -48.1636\n",
      "Epoch 96/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 389969.2188 - r_square: -36.9265 - val_loss: 422108.6875 - val_r_square: -48.1593\n",
      "Epoch 97/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 389934.1875 - r_square: -36.9230 - val_loss: 422072.0938 - val_r_square: -48.1551\n",
      "Epoch 98/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 389899.0938 - r_square: -36.9195 - val_loss: 422035.4688 - val_r_square: -48.1508\n",
      "Epoch 99/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 389864.2188 - r_square: -36.9159 - val_loss: 421998.8750 - val_r_square: -48.1465\n",
      "Epoch 100/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 389829.0312 - r_square: -36.9128 - val_loss: 421962.2500 - val_r_square: -48.1423\n",
      "Epoch 101/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 389793.8750 - r_square: -36.9090 - val_loss: 421925.5938 - val_r_square: -48.1380\n",
      "Epoch 102/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 389758.8125 - r_square: -36.9057 - val_loss: 421888.9375 - val_r_square: -48.1337\n",
      "Epoch 103/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 389723.6562 - r_square: -36.9022 - val_loss: 421852.3438 - val_r_square: -48.1295\n",
      "Epoch 104/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 389688.5312 - r_square: -36.8992 - val_loss: 421815.6562 - val_r_square: -48.1252\n",
      "Epoch 105/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 389653.4062 - r_square: -36.8957 - val_loss: 421778.9688 - val_r_square: -48.1209\n",
      "Epoch 106/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 389618.2812 - r_square: -36.8921 - val_loss: 421742.3125 - val_r_square: -48.1167\n",
      "Epoch 107/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 389583.0938 - r_square: -36.8889 - val_loss: 421705.5625 - val_r_square: -48.1124\n",
      "Epoch 108/500\n",
      "3/3 [==============================] - 0s 105ms/step - loss: 389547.9062 - r_square: -36.8852 - val_loss: 421668.8438 - val_r_square: -48.1081\n",
      "Epoch 109/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 389512.7812 - r_square: -36.8819 - val_loss: 421632.0625 - val_r_square: -48.1038\n",
      "Epoch 110/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 389477.5312 - r_square: -36.8786 - val_loss: 421595.2812 - val_r_square: -48.0995\n",
      "Epoch 111/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 389442.2500 - r_square: -36.8748 - val_loss: 421558.4688 - val_r_square: -48.0952\n",
      "Epoch 112/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 389406.9375 - r_square: -36.8715 - val_loss: 421521.6562 - val_r_square: -48.0910\n",
      "Epoch 113/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 389371.6250 - r_square: -36.8680 - val_loss: 421484.8438 - val_r_square: -48.0867\n",
      "Epoch 114/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 389336.4688 - r_square: -36.8647 - val_loss: 421448.0312 - val_r_square: -48.0824\n",
      "Epoch 115/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 389301.1562 - r_square: -36.8612 - val_loss: 421411.1562 - val_r_square: -48.0781\n",
      "Epoch 116/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 389265.7812 - r_square: -36.8578 - val_loss: 421374.3438 - val_r_square: -48.0738\n",
      "Epoch 117/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 389230.4375 - r_square: -36.8546 - val_loss: 421337.5000 - val_r_square: -48.0695\n",
      "Epoch 118/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 389195.0938 - r_square: -36.8512 - val_loss: 421300.5625 - val_r_square: -48.0652\n",
      "Epoch 119/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 389159.7188 - r_square: -36.8477 - val_loss: 421263.7188 - val_r_square: -48.0609\n",
      "Epoch 120/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 389124.4062 - r_square: -36.8442 - val_loss: 421226.7812 - val_r_square: -48.0566\n",
      "Epoch 121/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 389089.1250 - r_square: -36.8407 - val_loss: 421189.8125 - val_r_square: -48.0523\n",
      "Epoch 122/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 389053.6250 - r_square: -36.8373 - val_loss: 421152.9375 - val_r_square: -48.0480\n",
      "Epoch 123/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 389018.2812 - r_square: -36.8338 - val_loss: 421115.9062 - val_r_square: -48.0437\n",
      "Epoch 124/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 388982.8438 - r_square: -36.8304 - val_loss: 421078.9062 - val_r_square: -48.0394\n",
      "Epoch 125/500\n",
      "3/3 [==============================] - 0s 169ms/step - loss: 388947.4375 - r_square: -36.8268 - val_loss: 421041.9688 - val_r_square: -48.0351\n",
      "Epoch 126/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388912.0000 - r_square: -36.8234 - val_loss: 421005.0312 - val_r_square: -48.0308\n",
      "Epoch 127/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388876.5312 - r_square: -36.8201 - val_loss: 420968.0312 - val_r_square: -48.0265\n",
      "Epoch 128/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 388841.1562 - r_square: -36.8165 - val_loss: 420931.0000 - val_r_square: -48.0222\n",
      "Epoch 129/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 388805.5938 - r_square: -36.8133 - val_loss: 420894.0000 - val_r_square: -48.0179\n",
      "Epoch 130/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388770.2500 - r_square: -36.8095 - val_loss: 420857.0000 - val_r_square: -48.0136\n",
      "Epoch 131/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 388734.7188 - r_square: -36.8063 - val_loss: 420819.9688 - val_r_square: -48.0092\n",
      "Epoch 132/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 388699.1562 - r_square: -36.8027 - val_loss: 420782.9062 - val_r_square: -48.0049\n",
      "Epoch 133/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388663.7188 - r_square: -36.7995 - val_loss: 420745.8125 - val_r_square: -48.0006\n",
      "Epoch 134/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 388628.1875 - r_square: -36.7955 - val_loss: 420708.7188 - val_r_square: -47.9963\n",
      "Epoch 135/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 388592.5938 - r_square: -36.7922 - val_loss: 420671.5938 - val_r_square: -47.9920\n",
      "Epoch 136/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388557.0000 - r_square: -36.7889 - val_loss: 420634.5000 - val_r_square: -47.9877\n",
      "Epoch 137/500\n",
      "3/3 [==============================] - 0s 106ms/step - loss: 388521.4375 - r_square: -36.7855 - val_loss: 420597.4062 - val_r_square: -47.9833\n",
      "Epoch 138/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388486.0000 - r_square: -36.7820 - val_loss: 420560.2500 - val_r_square: -47.9790\n",
      "Epoch 139/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 388450.3750 - r_square: -36.7787 - val_loss: 420523.0938 - val_r_square: -47.9747\n",
      "Epoch 140/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 388414.7500 - r_square: -36.7753 - val_loss: 420486.0000 - val_r_square: -47.9704\n",
      "Epoch 141/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 388379.1875 - r_square: -36.7713 - val_loss: 420448.8750 - val_r_square: -47.9660\n",
      "Epoch 142/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 388343.6250 - r_square: -36.7680 - val_loss: 420411.6875 - val_r_square: -47.9617\n",
      "Epoch 143/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 388308.0312 - r_square: -36.7648 - val_loss: 420374.5312 - val_r_square: -47.9574\n",
      "Epoch 144/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 388272.4062 - r_square: -36.7612 - val_loss: 420337.3125 - val_r_square: -47.9530\n",
      "Epoch 145/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388236.7188 - r_square: -36.7578 - val_loss: 420300.1875 - val_r_square: -47.9487\n",
      "Epoch 146/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 388201.1875 - r_square: -36.7544 - val_loss: 420263.0000 - val_r_square: -47.9444\n",
      "Epoch 147/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 388165.5938 - r_square: -36.7510 - val_loss: 420225.8125 - val_r_square: -47.9400\n",
      "Epoch 148/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388129.8750 - r_square: -36.7472 - val_loss: 420188.6875 - val_r_square: -47.9357\n",
      "Epoch 149/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388094.3125 - r_square: -36.7439 - val_loss: 420151.4375 - val_r_square: -47.9314\n",
      "Epoch 150/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 388058.5000 - r_square: -36.7406 - val_loss: 420114.2500 - val_r_square: -47.9271\n",
      "Epoch 151/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 388022.8125 - r_square: -36.7368 - val_loss: 420077.0312 - val_r_square: -47.9227\n",
      "Epoch 152/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 387987.2812 - r_square: -36.7336 - val_loss: 420039.8125 - val_r_square: -47.9184\n",
      "Epoch 153/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 387951.6562 - r_square: -36.7302 - val_loss: 420002.6250 - val_r_square: -47.9141\n",
      "Epoch 154/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 387915.9375 - r_square: -36.7266 - val_loss: 419965.2812 - val_r_square: -47.9097\n",
      "Epoch 155/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 387880.2188 - r_square: -36.7233 - val_loss: 419928.0312 - val_r_square: -47.9054\n",
      "Epoch 156/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 387844.5000 - r_square: -36.7194 - val_loss: 419890.7500 - val_r_square: -47.9010\n",
      "Epoch 157/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 387808.6875 - r_square: -36.7161 - val_loss: 419853.5000 - val_r_square: -47.8967\n",
      "Epoch 158/500\n",
      "3/3 [==============================] - 0s 181ms/step - loss: 387773.0938 - r_square: -36.7129 - val_loss: 419816.1250 - val_r_square: -47.8923\n",
      "Epoch 159/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 387737.2500 - r_square: -36.7093 - val_loss: 419778.9062 - val_r_square: -47.8880\n",
      "Epoch 160/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 387701.5312 - r_square: -36.7055 - val_loss: 419741.5938 - val_r_square: -47.8837\n",
      "Epoch 161/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 387665.8125 - r_square: -36.7019 - val_loss: 419704.2500 - val_r_square: -47.8793\n",
      "Epoch 162/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 387630.0938 - r_square: -36.6986 - val_loss: 419666.9062 - val_r_square: -47.8750\n",
      "Epoch 163/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 387594.0625 - r_square: -36.6951 - val_loss: 419629.5625 - val_r_square: -47.8706\n",
      "Epoch 164/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 387558.5000 - r_square: -36.6920 - val_loss: 419592.1250 - val_r_square: -47.8663\n",
      "Epoch 165/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 387522.5938 - r_square: -36.6884 - val_loss: 419554.8438 - val_r_square: -47.8619\n",
      "Epoch 166/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 387486.7812 - r_square: -36.6850 - val_loss: 419517.4375 - val_r_square: -47.8576\n",
      "Epoch 167/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 387451.0312 - r_square: -36.6812 - val_loss: 419480.0625 - val_r_square: -47.8532\n",
      "Epoch 168/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 387415.2812 - r_square: -36.6776 - val_loss: 419442.6875 - val_r_square: -47.8489\n",
      "Epoch 169/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 387379.3125 - r_square: -36.6743 - val_loss: 419405.3125 - val_r_square: -47.8445\n",
      "Epoch 170/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 387343.4688 - r_square: -36.6708 - val_loss: 419367.8750 - val_r_square: -47.8401\n",
      "Epoch 171/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 387307.7812 - r_square: -36.6676 - val_loss: 419330.4062 - val_r_square: -47.8358\n",
      "Epoch 172/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 387271.7188 - r_square: -36.6639 - val_loss: 419292.9062 - val_r_square: -47.8314\n",
      "Epoch 173/500\n",
      "3/3 [==============================] - 0s 107ms/step - loss: 387235.8125 - r_square: -36.6604 - val_loss: 419255.4375 - val_r_square: -47.8270\n",
      "Epoch 174/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 387199.8750 - r_square: -36.6569 - val_loss: 419217.8750 - val_r_square: -47.8227\n",
      "Epoch 175/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 387163.8438 - r_square: -36.6532 - val_loss: 419180.3750 - val_r_square: -47.8183\n",
      "Epoch 176/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 387127.8750 - r_square: -36.6499 - val_loss: 419142.8125 - val_r_square: -47.8139\n",
      "Epoch 177/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 387091.9688 - r_square: -36.6465 - val_loss: 419105.1875 - val_r_square: -47.8095\n",
      "Epoch 178/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 387055.7812 - r_square: -36.6431 - val_loss: 419067.5000 - val_r_square: -47.8052\n",
      "Epoch 179/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 387019.8125 - r_square: -36.6396 - val_loss: 419029.8438 - val_r_square: -47.8008\n",
      "Epoch 180/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 386983.6562 - r_square: -36.6358 - val_loss: 418992.0625 - val_r_square: -47.7964\n",
      "Epoch 181/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 386947.4688 - r_square: -36.6322 - val_loss: 418954.3125 - val_r_square: -47.7920\n",
      "Epoch 182/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 386911.3125 - r_square: -36.6286 - val_loss: 418916.5000 - val_r_square: -47.7876\n",
      "Epoch 183/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 386875.2188 - r_square: -36.6256 - val_loss: 418878.6250 - val_r_square: -47.7832\n",
      "Epoch 184/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 386839.0000 - r_square: -36.6218 - val_loss: 418840.7188 - val_r_square: -47.7787\n",
      "Epoch 185/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 386802.0625 - r_square: -36.6184 - val_loss: 418802.7500 - val_r_square: -47.7743\n",
      "Epoch 186/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 386765.8438 - r_square: -36.6145 - val_loss: 418764.7500 - val_r_square: -47.7699\n",
      "Epoch 187/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 386729.5625 - r_square: -36.6109 - val_loss: 418726.5938 - val_r_square: -47.7655\n",
      "Epoch 188/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 386693.0312 - r_square: -36.6076 - val_loss: 418688.3750 - val_r_square: -47.7610\n",
      "Epoch 189/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 386656.3438 - r_square: -36.6039 - val_loss: 418650.0625 - val_r_square: -47.7565\n",
      "Epoch 190/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 386619.5938 - r_square: -36.6004 - val_loss: 418611.6562 - val_r_square: -47.7521\n",
      "Epoch 191/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 386582.8750 - r_square: -36.5969 - val_loss: 418573.0000 - val_r_square: -47.7476\n",
      "Epoch 192/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 386545.0625 - r_square: -36.5934 - val_loss: 418534.0938 - val_r_square: -47.7430\n",
      "Epoch 193/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 386507.9062 - r_square: -36.5897 - val_loss: 418494.8438 - val_r_square: -47.7385\n",
      "Epoch 194/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 386470.2188 - r_square: -36.5862 - val_loss: 418454.9375 - val_r_square: -47.7338\n",
      "Epoch 195/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 386430.7500 - r_square: -36.5819 - val_loss: 418413.7188 - val_r_square: -47.7290\n",
      "Epoch 196/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 386389.9688 - r_square: -36.5782 - val_loss: 418368.0312 - val_r_square: -47.7237\n",
      "Epoch 197/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 386340.9062 - r_square: -36.5733 - val_loss: 418239.4688 - val_r_square: -47.7087\n",
      "Epoch 198/500\n",
      "3/3 [==============================] - 0s 164ms/step - loss: 391906.0000 - r_square: -37.1146 - val_loss: 418283.5938 - val_r_square: -47.7138\n",
      "Epoch 199/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 386191.8750 - r_square: -36.5587 - val_loss: 417412.5000 - val_r_square: -47.6124\n",
      "Epoch 200/500\n",
      "3/3 [==============================] - 0s 147ms/step - loss: 379578.0938 - r_square: -35.9153 - val_loss: 323779.5000 - val_r_square: -36.7076\n",
      "Epoch 201/500\n",
      "3/3 [==============================] - 0s 153ms/step - loss: 941406.1250 - r_square: -90.5562 - val_loss: 134769.1719 - val_r_square: -14.6950\n",
      "Epoch 202/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 253768.1719 - r_square: -23.6799 - val_loss: 412309.2812 - val_r_square: -47.0180\n",
      "Epoch 203/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 383134.2188 - r_square: -36.2616 - val_loss: 417766.8125 - val_r_square: -47.6536\n",
      "Epoch 204/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 385894.3438 - r_square: -36.5296 - val_loss: 418039.6875 - val_r_square: -47.6854\n",
      "Epoch 205/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 386060.9062 - r_square: -36.5460 - val_loss: 418075.3125 - val_r_square: -47.6895\n",
      "Epoch 206/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 386078.7812 - r_square: -36.5479 - val_loss: 418067.0000 - val_r_square: -47.6885\n",
      "Epoch 207/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 386066.1562 - r_square: -36.5468 - val_loss: 418043.9688 - val_r_square: -47.6859\n",
      "Epoch 208/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 386043.3750 - r_square: -36.5443 - val_loss: 418014.5938 - val_r_square: -47.6824\n",
      "Epoch 209/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 386011.8438 - r_square: -36.5413 - val_loss: 417980.0938 - val_r_square: -47.6784\n",
      "Epoch 210/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 385978.6562 - r_square: -36.5382 - val_loss: 417935.9062 - val_r_square: -47.6733\n",
      "Epoch 211/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 385931.3750 - r_square: -36.5337 - val_loss: 417875.4375 - val_r_square: -47.6662\n",
      "Epoch 212/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 385855.6562 - r_square: -36.5261 - val_loss: 417668.1562 - val_r_square: -47.6421\n",
      "Epoch 213/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 385319.0625 - r_square: -36.4741 - val_loss: 413780.1875 - val_r_square: -47.1893\n",
      "Epoch 214/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 365995.3125 - r_square: -34.5945 - val_loss: 295195.4375 - val_r_square: -33.3786\n",
      "Epoch 215/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 252369.6875 - r_square: -23.5440 - val_loss: 130399.8594 - val_r_square: -14.1861\n",
      "Epoch 216/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 115903.3594 - r_square: -10.2718 - val_loss: 69418.2266 - val_r_square: -7.0840\n",
      "Epoch 217/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 78274.3672 - r_square: -6.6121 - val_loss: 51575.5625 - val_r_square: -5.0060\n",
      "Epoch 218/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 70163.6719 - r_square: -5.8232 - val_loss: 55850.2266 - val_r_square: -5.5038\n",
      "Epoch 219/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 53699.3438 - r_square: -4.2219 - val_loss: 32518.1621 - val_r_square: -2.7864\n",
      "Epoch 220/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 45305.8086 - r_square: -3.4056 - val_loss: 38004.6406 - val_r_square: -3.4253\n",
      "Epoch 221/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 40209.2734 - r_square: -2.9099 - val_loss: 27911.5078 - val_r_square: -2.2499\n",
      "Epoch 222/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 38016.0742 - r_square: -2.6966 - val_loss: 24993.9492 - val_r_square: -1.9100\n",
      "Epoch 223/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 32395.6445 - r_square: -2.1500 - val_loss: 27183.5566 - val_r_square: -2.1650\n",
      "Epoch 224/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 30448.4961 - r_square: -1.9606 - val_loss: 21583.5938 - val_r_square: -1.5128\n",
      "Epoch 225/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 28734.0996 - r_square: -1.7938 - val_loss: 23488.6289 - val_r_square: -1.7347\n",
      "Epoch 226/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 27154.0527 - r_square: -1.6402 - val_loss: 18236.5254 - val_r_square: -1.1230\n",
      "Epoch 227/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 27340.8945 - r_square: -1.6583 - val_loss: 23545.2891 - val_r_square: -1.7413\n",
      "Epoch 228/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 23035.1914 - r_square: -1.2396 - val_loss: 15934.5430 - val_r_square: -0.8549\n",
      "Epoch 229/500\n",
      "3/3 [==============================] - 0s 108ms/step - loss: 20700.8262 - r_square: -1.0125 - val_loss: 14060.6416 - val_r_square: -0.6366\n",
      "Epoch 230/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 20754.3574 - r_square: -1.0177 - val_loss: 16364.0850 - val_r_square: -0.9049\n",
      "Epoch 231/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 19599.1660 - r_square: -0.9054 - val_loss: 12110.1309 - val_r_square: -0.4095\n",
      "Epoch 232/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 19190.7617 - r_square: -0.8656 - val_loss: 9589.5518 - val_r_square: -0.1159\n",
      "Epoch 233/500\n",
      "3/3 [==============================] - 0s 168ms/step - loss: 17958.0879 - r_square: -0.7458 - val_loss: 11685.6865 - val_r_square: -0.3600\n",
      "Epoch 234/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 15851.2529 - r_square: -0.5408 - val_loss: 11524.8574 - val_r_square: -0.3413\n",
      "Epoch 235/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 15233.8887 - r_square: -0.4808 - val_loss: 8283.4658 - val_r_square: 0.0362\n",
      "Epoch 236/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 14492.4775 - r_square: -0.4087 - val_loss: 11333.4033 - val_r_square: -0.3190\n",
      "Epoch 237/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 13973.1152 - r_square: -0.3582 - val_loss: 7848.3438 - val_r_square: 0.0869\n",
      "Epoch 238/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 13260.9180 - r_square: -0.2889 - val_loss: 9288.3223 - val_r_square: -0.0808\n",
      "Epoch 239/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 14155.7070 - r_square: -0.3759 - val_loss: 9054.0674 - val_r_square: -0.0535\n",
      "Epoch 240/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 13417.3438 - r_square: -0.3041 - val_loss: 7273.0166 - val_r_square: 0.1539\n",
      "Epoch 241/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 13140.8818 - r_square: -0.2772 - val_loss: 8259.4053 - val_r_square: 0.0390\n",
      "Epoch 242/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 13210.1504 - r_square: -0.2840 - val_loss: 8052.4902 - val_r_square: 0.0631\n",
      "Epoch 243/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 11850.5996 - r_square: -0.1518 - val_loss: 6991.3384 - val_r_square: 0.1867\n",
      "Epoch 244/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 11350.8896 - r_square: -0.1032 - val_loss: 6651.7358 - val_r_square: 0.2263\n",
      "Epoch 245/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 11761.0156 - r_square: -0.1430 - val_loss: 10252.7471 - val_r_square: -0.1931\n",
      "Epoch 246/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 12008.4756 - r_square: -0.1671 - val_loss: 6279.1641 - val_r_square: 0.2697\n",
      "Epoch 247/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 11754.2969 - r_square: -0.1424 - val_loss: 7447.2817 - val_r_square: 0.1336\n",
      "Epoch 248/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 12423.0000 - r_square: -0.2074 - val_loss: 9797.5635 - val_r_square: -0.1401\n",
      "Epoch 249/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 12888.5332 - r_square: -0.2527 - val_loss: 6286.5889 - val_r_square: 0.2688\n",
      "Epoch 250/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 13216.5840 - r_square: -0.2846 - val_loss: 8402.7881 - val_r_square: 0.0223\n",
      "Epoch 251/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 11831.1904 - r_square: -0.1498 - val_loss: 5777.2388 - val_r_square: 0.3281\n",
      "Epoch 252/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 11476.1465 - r_square: -0.1153 - val_loss: 7306.6533 - val_r_square: 0.1500\n",
      "Epoch 253/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 11899.9004 - r_square: -0.1565 - val_loss: 6144.4771 - val_r_square: 0.2853\n",
      "Epoch 254/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 11188.6221 - r_square: -0.0874 - val_loss: 7101.4443 - val_r_square: 0.1739\n",
      "Epoch 255/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 10981.5352 - r_square: -0.0672 - val_loss: 6102.6050 - val_r_square: 0.2902\n",
      "Epoch 256/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 10708.6689 - r_square: -0.0407 - val_loss: 5761.6392 - val_r_square: 0.3299\n",
      "Epoch 257/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 11482.2705 - r_square: -0.1159 - val_loss: 9089.2412 - val_r_square: -0.0576\n",
      "Epoch 258/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 11182.3037 - r_square: -0.0868 - val_loss: 5453.3677 - val_r_square: 0.3658\n",
      "Epoch 259/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 11334.0967 - r_square: -0.1015 - val_loss: 7833.0044 - val_r_square: 0.0887\n",
      "Epoch 260/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 11844.8418 - r_square: -0.1512 - val_loss: 5501.5801 - val_r_square: 0.3602\n",
      "Epoch 261/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 10093.6777 - r_square: 0.0191 - val_loss: 6405.0806 - val_r_square: 0.2550\n",
      "Epoch 262/500\n",
      "3/3 [==============================] - 0s 149ms/step - loss: 10111.2617 - r_square: 0.0174 - val_loss: 6211.4009 - val_r_square: 0.2776\n",
      "Epoch 263/500\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 10243.2422 - r_square: 0.0046 - val_loss: 6220.0356 - val_r_square: 0.2766\n",
      "Epoch 264/500\n",
      "3/3 [==============================] - 0s 109ms/step - loss: 9783.0957 - r_square: 0.0493 - val_loss: 5350.1621 - val_r_square: 0.3779\n",
      "Epoch 265/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 10657.5674 - r_square: -0.0357 - val_loss: 7885.3931 - val_r_square: 0.0826\n",
      "Epoch 266/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 11032.8340 - r_square: -0.0722 - val_loss: 5203.5728 - val_r_square: 0.3949\n",
      "Epoch 267/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 10329.1445 - r_square: -0.0038 - val_loss: 6370.6011 - val_r_square: 0.2590\n",
      "Epoch 268/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 10270.1982 - r_square: 0.0020 - val_loss: 5567.3862 - val_r_square: 0.3526\n",
      "Epoch 269/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 9648.3525 - r_square: 0.0624 - val_loss: 6514.9927 - val_r_square: 0.2422\n",
      "Epoch 270/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 9758.5615 - r_square: 0.0517 - val_loss: 5347.5503 - val_r_square: 0.3782\n",
      "Epoch 271/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 10742.2979 - r_square: -0.0439 - val_loss: 6352.4380 - val_r_square: 0.2611\n",
      "Epoch 272/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 9526.1982 - r_square: 0.0743 - val_loss: 5562.3257 - val_r_square: 0.3532\n",
      "Epoch 273/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 10344.2207 - r_square: -0.0052 - val_loss: 5480.0308 - val_r_square: 0.3627\n",
      "Epoch 274/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 10161.4043 - r_square: 0.0125 - val_loss: 8095.7163 - val_r_square: 0.0581\n",
      "Epoch 275/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 9678.1309 - r_square: 0.0595 - val_loss: 5087.0200 - val_r_square: 0.4085\n",
      "Epoch 276/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 9832.4004 - r_square: 0.0445 - val_loss: 7256.5215 - val_r_square: 0.1558\n",
      "Epoch 277/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 9268.9404 - r_square: 0.0993 - val_loss: 5192.9209 - val_r_square: 0.3962\n",
      "Epoch 278/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 9938.1377 - r_square: 0.0343 - val_loss: 5434.7866 - val_r_square: 0.3680\n",
      "Epoch 279/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 10056.7100 - r_square: 0.0227 - val_loss: 5305.9009 - val_r_square: 0.3830\n",
      "Epoch 280/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 10847.9072 - r_square: -0.0542 - val_loss: 6662.1694 - val_r_square: 0.2251\n",
      "Epoch 281/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 9506.6094 - r_square: 0.0762 - val_loss: 4924.0952 - val_r_square: 0.4275\n",
      "Epoch 282/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 9734.6699 - r_square: 0.0541 - val_loss: 6361.5972 - val_r_square: 0.2601\n",
      "Epoch 283/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 8588.1963 - r_square: 0.1655 - val_loss: 4995.7017 - val_r_square: 0.4191\n",
      "Epoch 284/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 9703.4414 - r_square: 0.0571 - val_loss: 6961.4751 - val_r_square: 0.1902\n",
      "Epoch 285/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 9387.3633 - r_square: 0.0878 - val_loss: 4843.4570 - val_r_square: 0.4369\n",
      "Epoch 286/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 9870.3975 - r_square: 0.0408 - val_loss: 5780.7310 - val_r_square: 0.3277\n",
      "Epoch 287/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 9576.6484 - r_square: 0.0694 - val_loss: 5564.7578 - val_r_square: 0.3529\n",
      "Epoch 288/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 9588.7314 - r_square: 0.0683 - val_loss: 5030.6670 - val_r_square: 0.4151\n",
      "Epoch 289/500\n",
      "3/3 [==============================] - 0s 172ms/step - loss: 9506.7686 - r_square: 0.0762 - val_loss: 6019.6812 - val_r_square: 0.2999\n",
      "Epoch 290/500\n",
      "3/3 [==============================] - 0s 137ms/step - loss: 9255.8574 - r_square: 0.1006 - val_loss: 4814.8799 - val_r_square: 0.4402\n",
      "Epoch 291/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 9928.4043 - r_square: 0.0352 - val_loss: 8515.6055 - val_r_square: 0.0092\n",
      "Epoch 292/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 10450.8164 - r_square: -0.0156 - val_loss: 4871.1118 - val_r_square: 0.4337\n",
      "Epoch 293/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 9826.8643 - r_square: 0.0451 - val_loss: 7242.9639 - val_r_square: 0.1574\n",
      "Epoch 294/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 9430.2695 - r_square: 0.0837 - val_loss: 5277.6440 - val_r_square: 0.3863\n",
      "Epoch 295/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 9003.3379 - r_square: 0.1252 - val_loss: 5587.7021 - val_r_square: 0.3502\n",
      "Epoch 296/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 8934.8770 - r_square: 0.1318 - val_loss: 5304.5195 - val_r_square: 0.3832\n",
      "Epoch 297/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 9558.1133 - r_square: 0.0712 - val_loss: 5207.4111 - val_r_square: 0.3945\n",
      "Epoch 298/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 8481.6230 - r_square: 0.1759 - val_loss: 5205.0894 - val_r_square: 0.3948\n",
      "Epoch 299/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 9309.7617 - r_square: 0.0954 - val_loss: 4927.8071 - val_r_square: 0.4271\n",
      "Epoch 300/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 8735.5625 - r_square: 0.1512 - val_loss: 5986.3306 - val_r_square: 0.3038\n",
      "Epoch 301/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 9326.2021 - r_square: 0.0938 - val_loss: 4959.4165 - val_r_square: 0.4234\n",
      "Epoch 302/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 9204.6982 - r_square: 0.1056 - val_loss: 5763.8159 - val_r_square: 0.3297\n",
      "Epoch 303/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 8967.2246 - r_square: 0.1287 - val_loss: 4697.9028 - val_r_square: 0.4538\n",
      "Epoch 304/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 9119.8213 - r_square: 0.1138 - val_loss: 6549.7124 - val_r_square: 0.2382\n",
      "Epoch 305/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 9230.8047 - r_square: 0.1031 - val_loss: 5030.9233 - val_r_square: 0.4150\n",
      "Epoch 306/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 8594.2900 - r_square: 0.1650 - val_loss: 5024.1470 - val_r_square: 0.4158\n",
      "Epoch 307/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 8352.7168 - r_square: 0.1885 - val_loss: 5220.1768 - val_r_square: 0.3930\n",
      "Epoch 308/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 9508.2168 - r_square: 0.0761 - val_loss: 4742.9824 - val_r_square: 0.4486\n",
      "Epoch 309/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 9026.6514 - r_square: 0.1229 - val_loss: 6148.4209 - val_r_square: 0.2849\n",
      "Epoch 310/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 8964.4678 - r_square: 0.1290 - val_loss: 4812.5713 - val_r_square: 0.4405\n",
      "Epoch 311/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 8213.9404 - r_square: 0.2020 - val_loss: 5791.3770 - val_r_square: 0.3265\n",
      "Epoch 312/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 8981.8516 - r_square: 0.1273 - val_loss: 4695.9194 - val_r_square: 0.4541\n",
      "Epoch 313/500\n",
      "3/3 [==============================] - 0s 170ms/step - loss: 8106.4897 - r_square: 0.2124 - val_loss: 5418.8892 - val_r_square: 0.3699\n",
      "Epoch 314/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 8914.1602 - r_square: 0.1339 - val_loss: 5197.8252 - val_r_square: 0.3956\n",
      "Epoch 315/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 7801.5234 - r_square: 0.2421 - val_loss: 5081.3315 - val_r_square: 0.4092\n",
      "Epoch 316/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 8551.5537 - r_square: 0.1691 - val_loss: 4595.6841 - val_r_square: 0.4657\n",
      "Epoch 317/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 8884.9775 - r_square: 0.1367 - val_loss: 6043.6006 - val_r_square: 0.2971\n",
      "Epoch 318/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 8941.8955 - r_square: 0.1312 - val_loss: 4621.5264 - val_r_square: 0.4627\n",
      "Epoch 319/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 8396.9902 - r_square: 0.1842 - val_loss: 4897.5269 - val_r_square: 0.4306\n",
      "Epoch 320/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 8781.1045 - r_square: 0.1468 - val_loss: 4805.1719 - val_r_square: 0.4413\n",
      "Epoch 321/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 8567.8232 - r_square: 0.1675 - val_loss: 5484.3765 - val_r_square: 0.3622\n",
      "Epoch 322/500\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 8423.7041 - r_square: 0.1816 - val_loss: 4624.9106 - val_r_square: 0.4623\n",
      "Epoch 323/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 9132.8047 - r_square: 0.1126 - val_loss: 5538.9702 - val_r_square: 0.3559\n",
      "Epoch 324/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8280.0908 - r_square: 0.1955 - val_loss: 4554.6821 - val_r_square: 0.4705\n",
      "Epoch 325/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 8590.6865 - r_square: 0.1653 - val_loss: 5334.0776 - val_r_square: 0.3797\n",
      "Epoch 326/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 8642.0752 - r_square: 0.1603 - val_loss: 4838.8291 - val_r_square: 0.4374\n",
      "Epoch 327/500\n",
      "3/3 [==============================] - 0s 141ms/step - loss: 8229.5889 - r_square: 0.2004 - val_loss: 4516.2031 - val_r_square: 0.4750\n",
      "Epoch 328/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 8775.1025 - r_square: 0.1474 - val_loss: 5301.9517 - val_r_square: 0.3835\n",
      "Epoch 329/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 8285.0986 - r_square: 0.1950 - val_loss: 4685.6675 - val_r_square: 0.4553\n",
      "Epoch 330/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 8105.8242 - r_square: 0.2125 - val_loss: 4665.5327 - val_r_square: 0.4576\n",
      "Epoch 331/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8454.4238 - r_square: 0.1786 - val_loss: 6032.8413 - val_r_square: 0.2984\n",
      "Epoch 332/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8424.6602 - r_square: 0.1815 - val_loss: 4355.7134 - val_r_square: 0.4937\n",
      "Epoch 333/500\n",
      "3/3 [==============================] - 0s 171ms/step - loss: 8215.3242 - r_square: 0.2018 - val_loss: 5595.8008 - val_r_square: 0.3493\n",
      "Epoch 334/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8436.1699 - r_square: 0.1803 - val_loss: 4352.2314 - val_r_square: 0.4941\n",
      "Epoch 335/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 8573.5068 - r_square: 0.1670 - val_loss: 5196.5352 - val_r_square: 0.3958\n",
      "Epoch 336/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 7483.4912 - r_square: 0.2730 - val_loss: 4939.4233 - val_r_square: 0.4257\n",
      "Epoch 337/500\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 8160.5708 - r_square: 0.2072 - val_loss: 4536.2251 - val_r_square: 0.4727\n",
      "Epoch 338/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 7716.7236 - r_square: 0.2503 - val_loss: 5168.1436 - val_r_square: 0.3991\n",
      "Epoch 339/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 7907.4644 - r_square: 0.2318 - val_loss: 4542.4517 - val_r_square: 0.4719\n",
      "Epoch 340/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 8264.5469 - r_square: 0.1970 - val_loss: 5164.9922 - val_r_square: 0.3994\n",
      "Epoch 341/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 8005.5259 - r_square: 0.2222 - val_loss: 4452.3672 - val_r_square: 0.4824\n",
      "Epoch 342/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 8471.3496 - r_square: 0.1769 - val_loss: 5587.4482 - val_r_square: 0.3502\n",
      "Epoch 343/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 7554.8530 - r_square: 0.2661 - val_loss: 4539.2124 - val_r_square: 0.4723\n",
      "Epoch 344/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 7555.7676 - r_square: 0.2660 - val_loss: 5022.9248 - val_r_square: 0.4160\n",
      "Epoch 345/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 7799.8491 - r_square: 0.2422 - val_loss: 4320.6616 - val_r_square: 0.4978\n",
      "Epoch 346/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 7659.7456 - r_square: 0.2559 - val_loss: 5308.2920 - val_r_square: 0.3828\n",
      "Epoch 347/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 8284.4980 - r_square: 0.1951 - val_loss: 4369.0669 - val_r_square: 0.4921\n",
      "Epoch 348/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 7970.5806 - r_square: 0.2256 - val_loss: 4942.0898 - val_r_square: 0.4254\n",
      "Epoch 349/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 8515.3789 - r_square: 0.1726 - val_loss: 5173.4277 - val_r_square: 0.3985\n",
      "Epoch 350/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 7679.7510 - r_square: 0.2539 - val_loss: 4191.2607 - val_r_square: 0.5129\n",
      "Epoch 351/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 8353.8467 - r_square: 0.1884 - val_loss: 6504.7793 - val_r_square: 0.2434\n",
      "Epoch 352/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 9016.1562 - r_square: 0.1240 - val_loss: 4124.9888 - val_r_square: 0.5206\n",
      "Epoch 353/500\n",
      "3/3 [==============================] - 0s 174ms/step - loss: 7907.6196 - r_square: 0.2318 - val_loss: 5281.7461 - val_r_square: 0.3859\n",
      "Epoch 354/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 8407.9287 - r_square: 0.1831 - val_loss: 4208.5923 - val_r_square: 0.5108\n",
      "Epoch 355/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 8084.9639 - r_square: 0.2145 - val_loss: 4852.2534 - val_r_square: 0.4359\n",
      "Epoch 356/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 8484.3047 - r_square: 0.1757 - val_loss: 5126.8511 - val_r_square: 0.4039\n",
      "Epoch 357/500\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 8227.4834 - r_square: 0.2006 - val_loss: 4210.3452 - val_r_square: 0.5106\n",
      "Epoch 358/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 8489.6523 - r_square: 0.1752 - val_loss: 4811.9648 - val_r_square: 0.4406\n",
      "Epoch 359/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8009.2720 - r_square: 0.2219 - val_loss: 4850.7446 - val_r_square: 0.4361\n",
      "Epoch 360/500\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 7577.0454 - r_square: 0.2639 - val_loss: 4181.4146 - val_r_square: 0.5140\n",
      "Epoch 361/500\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 7972.1812 - r_square: 0.2255 - val_loss: 5025.9062 - val_r_square: 0.4157\n",
      "Epoch 362/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 8168.3970 - r_square: 0.2064 - val_loss: 4046.2598 - val_r_square: 0.5297\n",
      "Epoch 363/500\n",
      "3/3 [==============================] - 0s 125ms/step - loss: 8691.1982 - r_square: 0.1556 - val_loss: 4580.0522 - val_r_square: 0.4676\n",
      "Epoch 364/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 8203.1758 - r_square: 0.2030 - val_loss: 4837.8276 - val_r_square: 0.4376\n",
      "Epoch 365/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 7871.3340 - r_square: 0.2353 - val_loss: 4479.2520 - val_r_square: 0.4793\n",
      "Epoch 366/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 7831.9351 - r_square: 0.2391 - val_loss: 4353.3296 - val_r_square: 0.4940\n",
      "Epoch 367/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 7730.5029 - r_square: 0.2490 - val_loss: 4493.5894 - val_r_square: 0.4777\n",
      "Epoch 368/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8393.0537 - r_square: 0.1846 - val_loss: 4348.6021 - val_r_square: 0.4945\n",
      "Epoch 369/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 8020.8364 - r_square: 0.2208 - val_loss: 4719.7236 - val_r_square: 0.4513\n",
      "Epoch 370/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 7361.8340 - r_square: 0.2848 - val_loss: 4414.0015 - val_r_square: 0.4869\n",
      "Epoch 371/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 7833.3315 - r_square: 0.2390 - val_loss: 4445.9961 - val_r_square: 0.4832\n",
      "Epoch 372/500\n",
      "3/3 [==============================] - 0s 181ms/step - loss: 7779.0884 - r_square: 0.2443 - val_loss: 4417.8726 - val_r_square: 0.4865\n",
      "Epoch 373/500\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 8137.9673 - r_square: 0.2094 - val_loss: 4056.4819 - val_r_square: 0.5286\n",
      "Epoch 374/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 8180.9424 - r_square: 0.2052 - val_loss: 5104.0767 - val_r_square: 0.4066\n",
      "Epoch 375/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7712.5576 - r_square: 0.2507 - val_loss: 4022.6685 - val_r_square: 0.5325\n",
      "Epoch 376/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 7788.2827 - r_square: 0.2434 - val_loss: 5198.6465 - val_r_square: 0.3955\n",
      "Epoch 377/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7122.1216 - r_square: 0.3081 - val_loss: 4161.3916 - val_r_square: 0.5163\n",
      "Epoch 378/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 8275.6865 - r_square: 0.1960 - val_loss: 5282.2734 - val_r_square: 0.3858\n",
      "Epoch 379/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 7323.9688 - r_square: 0.2885 - val_loss: 4236.9058 - val_r_square: 0.5076\n",
      "Epoch 380/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 8384.7705 - r_square: 0.1854 - val_loss: 5322.7202 - val_r_square: 0.3811\n",
      "Epoch 381/500\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 7896.6616 - r_square: 0.2328 - val_loss: 4575.7222 - val_r_square: 0.4681\n",
      "Epoch 382/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 7975.4185 - r_square: 0.2252 - val_loss: 4329.0459 - val_r_square: 0.4968\n",
      "Epoch 383/500\n",
      "3/3 [==============================] - 0s 130ms/step - loss: 7793.6479 - r_square: 0.2429 - val_loss: 4150.8979 - val_r_square: 0.5176\n",
      "Epoch 384/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 7955.9819 - r_square: 0.2271 - val_loss: 6416.9380 - val_r_square: 0.2537\n",
      "Epoch 385/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 8136.5273 - r_square: 0.2095 - val_loss: 3873.6128 - val_r_square: 0.5499\n",
      "Epoch 386/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 7939.6035 - r_square: 0.2287 - val_loss: 5267.7412 - val_r_square: 0.3875\n",
      "Epoch 387/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8249.6924 - r_square: 0.1985 - val_loss: 4858.2949 - val_r_square: 0.4352\n",
      "Epoch 388/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 8203.7578 - r_square: 0.2030 - val_loss: 4020.4038 - val_r_square: 0.5328\n",
      "Epoch 389/500\n",
      "3/3 [==============================] - 0s 177ms/step - loss: 7835.5459 - r_square: 0.2388 - val_loss: 5199.9316 - val_r_square: 0.3954\n",
      "Epoch 390/500\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 7750.8345 - r_square: 0.2470 - val_loss: 4096.6763 - val_r_square: 0.5239\n",
      "Epoch 391/500\n",
      "3/3 [==============================] - 0s 110ms/step - loss: 7526.2031 - r_square: 0.2689 - val_loss: 4119.5317 - val_r_square: 0.5212\n",
      "Epoch 392/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7850.6392 - r_square: 0.2373 - val_loss: 5673.7251 - val_r_square: 0.3402\n",
      "Epoch 393/500\n",
      "3/3 [==============================] - 0s 123ms/step - loss: 7933.0938 - r_square: 0.2293 - val_loss: 3829.6218 - val_r_square: 0.5550\n",
      "Epoch 394/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 7974.3604 - r_square: 0.2253 - val_loss: 7949.9463 - val_r_square: 0.0751\n",
      "Epoch 395/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 8105.1636 - r_square: 0.2126 - val_loss: 3831.4021 - val_r_square: 0.5548\n",
      "Epoch 396/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8047.8164 - r_square: 0.2181 - val_loss: 6166.9116 - val_r_square: 0.2828\n",
      "Epoch 397/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 8483.3037 - r_square: 0.1758 - val_loss: 3859.1167 - val_r_square: 0.5516\n",
      "Epoch 398/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7246.8779 - r_square: 0.2960 - val_loss: 5694.5234 - val_r_square: 0.3378\n",
      "Epoch 399/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 7494.9795 - r_square: 0.2719 - val_loss: 3840.2520 - val_r_square: 0.5538\n",
      "Epoch 400/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7768.8833 - r_square: 0.2453 - val_loss: 6220.5464 - val_r_square: 0.2765\n",
      "Epoch 401/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 7748.0205 - r_square: 0.2473 - val_loss: 3820.8682 - val_r_square: 0.5560\n",
      "Epoch 402/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 8062.3896 - r_square: 0.2167 - val_loss: 5607.8228 - val_r_square: 0.3479\n",
      "Epoch 403/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 8305.1143 - r_square: 0.1931 - val_loss: 3809.6819 - val_r_square: 0.5573\n",
      "Epoch 404/500\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 7539.2725 - r_square: 0.2676 - val_loss: 5938.1733 - val_r_square: 0.3094\n",
      "Epoch 405/500\n",
      "3/3 [==============================] - 0s 153ms/step - loss: 7393.6704 - r_square: 0.2818 - val_loss: 3798.4761 - val_r_square: 0.5586\n",
      "Epoch 406/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 8405.7842 - r_square: 0.1833 - val_loss: 4495.0947 - val_r_square: 0.4775\n",
      "Epoch 407/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 7433.9336 - r_square: 0.2778 - val_loss: 4610.1489 - val_r_square: 0.4641\n",
      "Epoch 408/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7530.0508 - r_square: 0.2685 - val_loss: 3867.3889 - val_r_square: 0.5506\n",
      "Epoch 409/500\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 7388.7837 - r_square: 0.2822 - val_loss: 4797.0918 - val_r_square: 0.4423\n",
      "Epoch 410/500\n",
      "3/3 [==============================] - 0s 111ms/step - loss: 7412.3345 - r_square: 0.2800 - val_loss: 3790.3569 - val_r_square: 0.5596\n",
      "Epoch 411/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 8094.1128 - r_square: 0.2137 - val_loss: 5917.1426 - val_r_square: 0.3119\n",
      "Epoch 412/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 7449.8623 - r_square: 0.2763 - val_loss: 3727.7822 - val_r_square: 0.5669\n",
      "Epoch 413/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 7484.2002 - r_square: 0.2730 - val_loss: 6025.7407 - val_r_square: 0.2992\n",
      "Epoch 414/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 8386.5898 - r_square: 0.1852 - val_loss: 3792.9651 - val_r_square: 0.5593\n",
      "Epoch 415/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 8162.7349 - r_square: 0.2070 - val_loss: 7171.5278 - val_r_square: 0.1658\n",
      "Epoch 416/500\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 8342.6846 - r_square: 0.1895 - val_loss: 4201.6758 - val_r_square: 0.5117\n",
      "Epoch 417/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 8781.5781 - r_square: 0.1468 - val_loss: 6709.6943 - val_r_square: 0.2196\n",
      "Epoch 418/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7778.2964 - r_square: 0.2444 - val_loss: 3722.5225 - val_r_square: 0.5675\n",
      "Epoch 419/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 7809.4160 - r_square: 0.2413 - val_loss: 6943.7466 - val_r_square: 0.1923\n",
      "Epoch 420/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 8077.5044 - r_square: 0.2153 - val_loss: 3655.6040 - val_r_square: 0.5753\n",
      "Epoch 421/500\n",
      "3/3 [==============================] - 0s 171ms/step - loss: 7839.1299 - r_square: 0.2385 - val_loss: 5772.6685 - val_r_square: 0.3287\n",
      "Epoch 422/500\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 7653.2026 - r_square: 0.2565 - val_loss: 3666.4727 - val_r_square: 0.5740\n",
      "Epoch 423/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 6973.3755 - r_square: 0.3227 - val_loss: 5105.0264 - val_r_square: 0.4065\n",
      "Epoch 424/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 6985.0913 - r_square: 0.3215 - val_loss: 4034.6714 - val_r_square: 0.5311\n",
      "Epoch 425/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 7552.8813 - r_square: 0.2663 - val_loss: 4824.3042 - val_r_square: 0.4392\n",
      "Epoch 426/500\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 7659.4717 - r_square: 0.2559 - val_loss: 3777.1675 - val_r_square: 0.5611\n",
      "Epoch 427/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7690.1675 - r_square: 0.2529 - val_loss: 4404.8320 - val_r_square: 0.4880\n",
      "Epoch 428/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 7020.7036 - r_square: 0.3181 - val_loss: 3957.4346 - val_r_square: 0.5401\n",
      "Epoch 429/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7851.0405 - r_square: 0.2373 - val_loss: 4069.0825 - val_r_square: 0.5271\n",
      "Epoch 430/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7742.8667 - r_square: 0.2478 - val_loss: 3613.8132 - val_r_square: 0.5802\n",
      "Epoch 431/500\n",
      "3/3 [==============================] - 0s 112ms/step - loss: 7520.2905 - r_square: 0.2695 - val_loss: 5181.8398 - val_r_square: 0.3975\n",
      "Epoch 432/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 7712.4751 - r_square: 0.2508 - val_loss: 3631.3606 - val_r_square: 0.5781\n",
      "Epoch 433/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 7393.6284 - r_square: 0.2818 - val_loss: 4689.8706 - val_r_square: 0.4548\n",
      "Epoch 434/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 7079.0356 - r_square: 0.3124 - val_loss: 4196.4727 - val_r_square: 0.5123\n",
      "Epoch 435/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 7524.5742 - r_square: 0.2691 - val_loss: 3773.2407 - val_r_square: 0.5616\n",
      "Epoch 436/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 6937.5776 - r_square: 0.3261 - val_loss: 4611.5010 - val_r_square: 0.4640\n",
      "Epoch 437/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 6936.6865 - r_square: 0.3262 - val_loss: 3908.0015 - val_r_square: 0.5459\n",
      "Epoch 438/500\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 7348.8140 - r_square: 0.2862 - val_loss: 4462.1523 - val_r_square: 0.4814\n",
      "Epoch 439/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7573.2773 - r_square: 0.2643 - val_loss: 3559.3633 - val_r_square: 0.5865\n",
      "Epoch 440/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 7525.1978 - r_square: 0.2690 - val_loss: 5432.6641 - val_r_square: 0.3683\n",
      "Epoch 441/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 7332.3491 - r_square: 0.2878 - val_loss: 3577.1726 - val_r_square: 0.5844\n",
      "Epoch 442/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 7320.3965 - r_square: 0.2889 - val_loss: 4784.0132 - val_r_square: 0.4439\n",
      "Epoch 443/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7103.3501 - r_square: 0.3100 - val_loss: 3607.1704 - val_r_square: 0.5809\n",
      "Epoch 444/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 6925.3120 - r_square: 0.3273 - val_loss: 4372.2427 - val_r_square: 0.4918\n",
      "Epoch 445/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7655.6777 - r_square: 0.2563 - val_loss: 3881.8054 - val_r_square: 0.5490\n",
      "Epoch 446/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 6853.9775 - r_square: 0.3343 - val_loss: 4216.3677 - val_r_square: 0.5100\n",
      "Epoch 447/500\n",
      "3/3 [==============================] - 0s 121ms/step - loss: 7468.3135 - r_square: 0.2745 - val_loss: 3688.2104 - val_r_square: 0.5715\n",
      "Epoch 448/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 7133.2402 - r_square: 0.3071 - val_loss: 4017.2590 - val_r_square: 0.5332\n",
      "Epoch 449/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7100.1494 - r_square: 0.3103 - val_loss: 3818.3501 - val_r_square: 0.5563\n",
      "Epoch 450/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 7084.7725 - r_square: 0.3118 - val_loss: 4078.0535 - val_r_square: 0.5261\n",
      "Epoch 451/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 6700.9893 - r_square: 0.3492 - val_loss: 3658.3333 - val_r_square: 0.5750\n",
      "Epoch 452/500\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 7283.3491 - r_square: 0.2925 - val_loss: 4244.2407 - val_r_square: 0.5067\n",
      "Epoch 453/500\n",
      "3/3 [==============================] - 0s 160ms/step - loss: 7705.0806 - r_square: 0.2515 - val_loss: 3954.0110 - val_r_square: 0.5406\n",
      "Epoch 454/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 6741.0664 - r_square: 0.3453 - val_loss: 3937.1548 - val_r_square: 0.5425\n",
      "Epoch 455/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 6842.8511 - r_square: 0.3354 - val_loss: 4493.9141 - val_r_square: 0.4777\n",
      "Epoch 456/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 7424.1470 - r_square: 0.2788 - val_loss: 3642.5115 - val_r_square: 0.5768\n",
      "Epoch 457/500\n",
      "3/3 [==============================] - 0s 126ms/step - loss: 7183.8140 - r_square: 0.3022 - val_loss: 5109.9380 - val_r_square: 0.4059\n",
      "Epoch 458/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 6721.7461 - r_square: 0.3471 - val_loss: 3498.9788 - val_r_square: 0.5935\n",
      "Epoch 459/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7014.0537 - r_square: 0.3187 - val_loss: 4229.5015 - val_r_square: 0.5085\n",
      "Epoch 460/500\n",
      "3/3 [==============================] - 0s 114ms/step - loss: 7016.7129 - r_square: 0.3185 - val_loss: 3835.4299 - val_r_square: 0.5544\n",
      "Epoch 461/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7162.2129 - r_square: 0.3043 - val_loss: 4462.2710 - val_r_square: 0.4814\n",
      "Epoch 462/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 7116.2427 - r_square: 0.3088 - val_loss: 3449.1162 - val_r_square: 0.5994\n",
      "Epoch 463/500\n",
      "3/3 [==============================] - 0s 128ms/step - loss: 7100.8472 - r_square: 0.3103 - val_loss: 5186.4438 - val_r_square: 0.3970\n",
      "Epoch 464/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 6975.0962 - r_square: 0.3225 - val_loss: 3447.2415 - val_r_square: 0.5996\n",
      "Epoch 465/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7790.4937 - r_square: 0.2432 - val_loss: 4747.4116 - val_r_square: 0.4482\n",
      "Epoch 466/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7342.3999 - r_square: 0.2868 - val_loss: 3592.2231 - val_r_square: 0.5827\n",
      "Epoch 467/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 6877.5908 - r_square: 0.3320 - val_loss: 4061.2068 - val_r_square: 0.5281\n",
      "Epoch 468/500\n",
      "3/3 [==============================] - 0s 146ms/step - loss: 7083.6587 - r_square: 0.3120 - val_loss: 4317.0503 - val_r_square: 0.4983\n",
      "Epoch 469/500\n",
      "3/3 [==============================] - 0s 122ms/step - loss: 6504.8442 - r_square: 0.3683 - val_loss: 3359.7480 - val_r_square: 0.6098\n",
      "Epoch 470/500\n",
      "3/3 [==============================] - 0s 129ms/step - loss: 7237.3657 - r_square: 0.2970 - val_loss: 5468.7798 - val_r_square: 0.3642\n",
      "Epoch 471/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7479.4565 - r_square: 0.2735 - val_loss: 3576.7478 - val_r_square: 0.5845\n",
      "Epoch 472/500\n",
      "3/3 [==============================] - 0s 127ms/step - loss: 7215.0000 - r_square: 0.2992 - val_loss: 3718.9739 - val_r_square: 0.5679\n",
      "Epoch 473/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 6621.6475 - r_square: 0.3569 - val_loss: 4464.8999 - val_r_square: 0.4811\n",
      "Epoch 474/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 6985.6963 - r_square: 0.3215 - val_loss: 3516.5801 - val_r_square: 0.5915\n",
      "Epoch 475/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 7225.5610 - r_square: 0.2982 - val_loss: 4726.0107 - val_r_square: 0.4507\n",
      "Epoch 476/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 7024.2505 - r_square: 0.3177 - val_loss: 3530.1697 - val_r_square: 0.5899\n",
      "Epoch 477/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7639.9912 - r_square: 0.2579 - val_loss: 3925.1472 - val_r_square: 0.5439\n",
      "Epoch 478/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 6699.0928 - r_square: 0.3494 - val_loss: 3875.9304 - val_r_square: 0.5497\n",
      "Epoch 479/500\n",
      "3/3 [==============================] - 0s 115ms/step - loss: 6998.9043 - r_square: 0.3202 - val_loss: 3835.4233 - val_r_square: 0.5544\n",
      "Epoch 480/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 6677.4536 - r_square: 0.3515 - val_loss: 3453.1025 - val_r_square: 0.5989\n",
      "Epoch 481/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 6652.3164 - r_square: 0.3539 - val_loss: 4252.5034 - val_r_square: 0.5058\n",
      "Epoch 482/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7111.5327 - r_square: 0.3093 - val_loss: 3398.8694 - val_r_square: 0.6052\n",
      "Epoch 483/500\n",
      "3/3 [==============================] - 0s 134ms/step - loss: 7437.0410 - r_square: 0.2776 - val_loss: 3864.8447 - val_r_square: 0.5510\n",
      "Epoch 484/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 6553.2690 - r_square: 0.3635 - val_loss: 4168.0806 - val_r_square: 0.5156\n",
      "Epoch 485/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 6852.8208 - r_square: 0.3344 - val_loss: 3226.5737 - val_r_square: 0.6253\n",
      "Epoch 486/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7150.1963 - r_square: 0.3055 - val_loss: 5620.5181 - val_r_square: 0.3465\n",
      "Epoch 487/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 7253.6152 - r_square: 0.2954 - val_loss: 3245.2954 - val_r_square: 0.6231\n",
      "Epoch 488/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 7311.6934 - r_square: 0.2898 - val_loss: 4100.4600 - val_r_square: 0.5235\n",
      "Epoch 489/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7178.8389 - r_square: 0.3027 - val_loss: 3634.1292 - val_r_square: 0.5778\n",
      "Epoch 490/500\n",
      "3/3 [==============================] - 0s 124ms/step - loss: 6602.5742 - r_square: 0.3588 - val_loss: 3979.6831 - val_r_square: 0.5376\n",
      "Epoch 491/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 6577.3667 - r_square: 0.3612 - val_loss: 3426.1992 - val_r_square: 0.6020\n",
      "Epoch 492/500\n",
      "3/3 [==============================] - 0s 113ms/step - loss: 7183.3320 - r_square: 0.3023 - val_loss: 3360.4346 - val_r_square: 0.6097\n",
      "Epoch 493/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 6889.0020 - r_square: 0.3309 - val_loss: 5811.7583 - val_r_square: 0.3242\n",
      "Epoch 494/500\n",
      "3/3 [==============================] - 0s 116ms/step - loss: 6919.1914 - r_square: 0.3280 - val_loss: 3369.2336 - val_r_square: 0.6087\n",
      "Epoch 495/500\n",
      "3/3 [==============================] - 0s 119ms/step - loss: 7484.5981 - r_square: 0.2730 - val_loss: 6894.8560 - val_r_square: 0.1981\n",
      "Epoch 496/500\n",
      "3/3 [==============================] - 0s 117ms/step - loss: 7907.4521 - r_square: 0.2318 - val_loss: 3258.6785 - val_r_square: 0.6216\n",
      "Epoch 497/500\n",
      "3/3 [==============================] - 0s 118ms/step - loss: 7731.3599 - r_square: 0.2490 - val_loss: 5041.7407 - val_r_square: 0.4139\n",
      "Epoch 498/500\n",
      "3/3 [==============================] - 0s 177ms/step - loss: 7744.0249 - r_square: 0.2477 - val_loss: 3442.1853 - val_r_square: 0.6002\n",
      "Epoch 499/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7410.3101 - r_square: 0.2802 - val_loss: 3500.3110 - val_r_square: 0.5934\n",
      "Epoch 500/500\n",
      "3/3 [==============================] - 0s 120ms/step - loss: 7470.9849 - r_square: 0.2743 - val_loss: 4219.1016 - val_r_square: 0.5097\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras import regularizers\n",
    "\n",
    "model_bd2 = Sequential()\n",
    "model_bd2.add(Bidirectional(LSTM(64, activation='relu', return_sequences=True, kernel_regularizer=regularizers.l2(0.01)), input_shape=(10, 5)))\n",
    "model_bd2.add(Dropout(0.2))\n",
    "model_bd2.add(Bidirectional(LSTM(64, activation='relu', kernel_regularizer=regularizers.l2(0.01))))\n",
    "model_bd2.add(Dropout(0.2))\n",
    "model_bd2.add(Dense(1, activation='linear'))\n",
    "\n",
    "model_bd2.compile(loss='mean_squared_error', optimizer=tf.keras.optimizers.Adam(learning_rate=0.01), metrics=[tfa.metrics.RSquare()])\n",
    "\n",
    "\n",
    "history_bd2 = model_bd2.fit(X_train_rs, y_train1, epochs=500, batch_size=350, validation_data=(X_val_rs, y_val1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 382,
   "id": "ef924b41-4a6d-47c5-81a8-258c9873cb87",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 1s 7ms/step\n",
      "59.506121402854866\n"
     ]
    }
   ],
   "source": [
    "y_pred_bd2_tr = model_bd2.predict(X_train_rs)\n",
    "bd2_rmse_train = np.sqrt(mean_squared_error(y_train1, y_pred_bd2_tr))\n",
    "print(bd2_rmse_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40ef200f-9362-4bc0-89c9-395f541e84e3",
   "metadata": {},
   "source": [
    "## ------------ GRU (Gated Recurrent Unit) ----------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 383,
   "id": "308e0691-e9c8-4cf7-9547-95af43be0080",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/400\n",
      "8/8 [==============================] - 3s 69ms/step - loss: 2215788800.0000 - r_square: -215495.7812 - val_loss: 425337.4375 - val_r_square: -48.5353\n",
      "Epoch 2/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 393536.6875 - r_square: -37.2736 - val_loss: 426554.5625 - val_r_square: -48.6771\n",
      "Epoch 3/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 394395.7812 - r_square: -37.3569 - val_loss: 426915.3750 - val_r_square: -48.7191\n",
      "Epoch 4/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 393988.1875 - r_square: -37.3176 - val_loss: 421965.5000 - val_r_square: -48.1426\n",
      "Epoch 5/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 335288.5000 - r_square: -31.6085 - val_loss: 190813.0781 - val_r_square: -21.2223\n",
      "Epoch 6/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 158354.8438 - r_square: -14.4008 - val_loss: 158651.4219 - val_r_square: -17.4767\n",
      "Epoch 7/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 133180.6875 - r_square: -11.9526 - val_loss: 82726.2344 - val_r_square: -8.6344\n",
      "Epoch 8/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 108846.6797 - r_square: -9.5859 - val_loss: 78229.8359 - val_r_square: -8.1107\n",
      "Epoch 9/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 83685.2656 - r_square: -7.1389 - val_loss: 56512.3086 - val_r_square: -5.5815\n",
      "Epoch 10/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 71256.6016 - r_square: -5.9301 - val_loss: 54335.3867 - val_r_square: -5.3280\n",
      "Epoch 11/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 72509.0078 - r_square: -6.0519 - val_loss: 51491.6602 - val_r_square: -4.9968\n",
      "Epoch 12/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 66714.4297 - r_square: -5.4883 - val_loss: 63646.7070 - val_r_square: -6.4124\n",
      "Epoch 13/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 260836163584.0000 - r_square: -25367648.0000 - val_loss: 202271.7500 - val_r_square: -22.5568\n",
      "Epoch 14/400\n",
      "8/8 [==============================] - 0s 23ms/step - loss: 289413.3125 - r_square: -27.1469 - val_loss: 420385.7500 - val_r_square: -47.9586\n",
      "Epoch 15/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 389632.4062 - r_square: -36.8937 - val_loss: 428609.4062 - val_r_square: -48.9164\n",
      "Epoch 16/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 390092.9688 - r_square: -36.9388 - val_loss: 424147.6875 - val_r_square: -48.3968\n",
      "Epoch 17/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 386182.6875 - r_square: -36.5582 - val_loss: 417841.5000 - val_r_square: -47.6623\n",
      "Epoch 18/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 381687.9375 - r_square: -36.1214 - val_loss: 412849.1250 - val_r_square: -47.0809\n",
      "Epoch 19/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 376971.7500 - r_square: -35.6624 - val_loss: 405896.4688 - val_r_square: -46.2712\n",
      "Epoch 20/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 370626.5000 - r_square: -35.0453 - val_loss: 397674.6875 - val_r_square: -45.3137\n",
      "Epoch 21/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 362898.6250 - r_square: -34.2940 - val_loss: 386884.4688 - val_r_square: -44.0570\n",
      "Epoch 22/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 352179.0625 - r_square: -33.2514 - val_loss: 371144.5312 - val_r_square: -42.2239\n",
      "Epoch 23/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 332645.1562 - r_square: -31.3517 - val_loss: 357182.8438 - val_r_square: -40.5979\n",
      "Epoch 24/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 324938.1875 - r_square: -30.6019 - val_loss: 344798.0312 - val_r_square: -39.1556\n",
      "Epoch 25/400\n",
      "8/8 [==============================] - 0s 24ms/step - loss: 312060.5312 - r_square: -29.3496 - val_loss: 334769.2500 - val_r_square: -37.9876\n",
      "Epoch 26/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 316021.7188 - r_square: -29.7349 - val_loss: 333598.9375 - val_r_square: -37.8513\n",
      "Epoch 27/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 310890.7500 - r_square: -29.2359 - val_loss: 331210.1875 - val_r_square: -37.5731\n",
      "Epoch 28/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 309847.9062 - r_square: -29.1344 - val_loss: 327697.9062 - val_r_square: -37.1641\n",
      "Epoch 29/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 305903.7188 - r_square: -28.7508 - val_loss: 322639.3125 - val_r_square: -36.5750\n",
      "Epoch 30/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 300753.7188 - r_square: -28.2498 - val_loss: 317507.2812 - val_r_square: -35.9773\n",
      "Epoch 31/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 295764.6250 - r_square: -27.7648 - val_loss: 312902.4375 - val_r_square: -35.4410\n",
      "Epoch 32/400\n",
      "8/8 [==============================] - 0s 22ms/step - loss: 291371.3438 - r_square: -27.3379 - val_loss: 308031.2812 - val_r_square: -34.8737\n",
      "Epoch 33/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 286170.9062 - r_square: -26.8317 - val_loss: 302675.1562 - val_r_square: -34.2499\n",
      "Epoch 34/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 280777.9062 - r_square: -26.3072 - val_loss: 296822.5312 - val_r_square: -33.5683\n",
      "Epoch 35/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 275197.2812 - r_square: -25.7644 - val_loss: 291334.2812 - val_r_square: -32.9291\n",
      "Epoch 36/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 270104.4062 - r_square: -25.2691 - val_loss: 286127.2188 - val_r_square: -32.3227\n",
      "Epoch 37/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 266349.5625 - r_square: -24.9038 - val_loss: 282264.5000 - val_r_square: -31.8729\n",
      "Epoch 38/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 264644.0312 - r_square: -24.7378 - val_loss: 282604.3125 - val_r_square: -31.9124\n",
      "Epoch 39/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 262791.6875 - r_square: -24.5579 - val_loss: 277933.1562 - val_r_square: -31.3684\n",
      "Epoch 40/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 254068.1875 - r_square: -23.7095 - val_loss: 270537.0000 - val_r_square: -30.5071\n",
      "Epoch 41/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 247066.9219 - r_square: -23.0284 - val_loss: 261659.8125 - val_r_square: -29.4732\n",
      "Epoch 42/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 239778.6719 - r_square: -22.3196 - val_loss: 257355.8438 - val_r_square: -28.9720\n",
      "Epoch 43/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 234635.5469 - r_square: -21.8196 - val_loss: 252363.5938 - val_r_square: -28.3906\n",
      "Epoch 44/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 229459.9531 - r_square: -21.3163 - val_loss: 249600.7344 - val_r_square: -28.0688\n",
      "Epoch 45/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 223804.1406 - r_square: -20.7660 - val_loss: 242046.9062 - val_r_square: -27.1891\n",
      "Epoch 46/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 220544.7188 - r_square: -20.4495 - val_loss: 253378.1719 - val_r_square: -28.5087\n",
      "Epoch 47/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 223274.2344 - r_square: -20.7145 - val_loss: 247375.4375 - val_r_square: -27.8096\n",
      "Epoch 48/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 216003.0156 - r_square: -20.0073 - val_loss: 236494.8438 - val_r_square: -26.5425\n",
      "Epoch 49/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 206731.4375 - r_square: -19.1056 - val_loss: 221516.0156 - val_r_square: -24.7980\n",
      "Epoch 50/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 198592.3906 - r_square: -18.3141 - val_loss: 212176.4531 - val_r_square: -23.7103\n",
      "Epoch 51/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 198795.7188 - r_square: -18.3340 - val_loss: 217334.3281 - val_r_square: -24.3110\n",
      "Epoch 52/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 200658.9688 - r_square: -18.5150 - val_loss: 215980.7500 - val_r_square: -24.1534\n",
      "Epoch 53/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 190174.6719 - r_square: -17.4956 - val_loss: 204029.2812 - val_r_square: -22.7615\n",
      "Epoch 54/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 184734.8594 - r_square: -16.9665 - val_loss: 197828.0156 - val_r_square: -22.0393\n",
      "Epoch 55/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 177837.9062 - r_square: -16.2958 - val_loss: 194776.4531 - val_r_square: -21.6839\n",
      "Epoch 56/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 171319.3438 - r_square: -15.6618 - val_loss: 186160.4219 - val_r_square: -20.6805\n",
      "Epoch 57/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 166084.6250 - r_square: -15.1526 - val_loss: 181999.1094 - val_r_square: -20.1958\n",
      "Epoch 58/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 171547.5625 - r_square: -15.6839 - val_loss: 188046.9219 - val_r_square: -20.9002\n",
      "Epoch 59/400\n",
      "8/8 [==============================] - 0s 22ms/step - loss: 171460.1562 - r_square: -15.6754 - val_loss: 185009.7812 - val_r_square: -20.5465\n",
      "Epoch 60/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 165282.2188 - r_square: -15.0749 - val_loss: 185517.3281 - val_r_square: -20.6056\n",
      "Epoch 61/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 161799.7656 - r_square: -14.7360 - val_loss: 167897.6094 - val_r_square: -18.5536\n",
      "Epoch 62/400\n",
      "8/8 [==============================] - 0s 26ms/step - loss: 152819.2188 - r_square: -13.8626 - val_loss: 162950.2188 - val_r_square: -17.9774\n",
      "Epoch 63/400\n",
      "8/8 [==============================] - 0s 25ms/step - loss: 152669.6719 - r_square: -13.8480 - val_loss: 163170.8125 - val_r_square: -18.0031\n",
      "Epoch 64/400\n",
      "8/8 [==============================] - 0s 22ms/step - loss: 162608.0625 - r_square: -14.8145 - val_loss: 191231.0625 - val_r_square: -21.2710\n",
      "Epoch 65/400\n",
      "8/8 [==============================] - 0s 25ms/step - loss: 198194.1562 - r_square: -18.2755 - val_loss: 226199.7031 - val_r_square: -25.3435\n",
      "Epoch 66/400\n",
      "8/8 [==============================] - 0s 22ms/step - loss: 203853.9062 - r_square: -18.8258 - val_loss: 226970.0938 - val_r_square: -25.4332\n",
      "Epoch 67/400\n",
      "8/8 [==============================] - 0s 25ms/step - loss: 193050.5781 - r_square: -17.7753 - val_loss: 218470.0156 - val_r_square: -24.4433\n",
      "Epoch 68/400\n",
      "8/8 [==============================] - 0s 25ms/step - loss: 187583.1406 - r_square: -17.2435 - val_loss: 206346.9375 - val_r_square: -23.0314\n",
      "Epoch 69/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 180723.6875 - r_square: -16.5762 - val_loss: 200285.5469 - val_r_square: -22.3255\n",
      "Epoch 70/400\n",
      "8/8 [==============================] - 0s 22ms/step - loss: 180998.9844 - r_square: -16.6030 - val_loss: 193143.1719 - val_r_square: -21.4937\n",
      "Epoch 71/400\n",
      "8/8 [==============================] - 0s 22ms/step - loss: 172348.0625 - r_square: -15.7618 - val_loss: 169930.7031 - val_r_square: -18.7903\n",
      "Epoch 72/400\n",
      "8/8 [==============================] - 0s 26ms/step - loss: 138016.9375 - r_square: -12.4229 - val_loss: 132140.9844 - val_r_square: -14.3893\n",
      "Epoch 73/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 115663.3594 - r_square: -10.2488 - val_loss: 98109.5000 - val_r_square: -10.4260\n",
      "Epoch 74/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 81496.3438 - r_square: -6.9260 - val_loss: 67003.6406 - val_r_square: -6.8033\n",
      "Epoch 75/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 65961.6797 - r_square: -5.4151 - val_loss: 53359.1172 - val_r_square: -5.2143\n",
      "Epoch 76/400\n",
      "8/8 [==============================] - 0s 23ms/step - loss: 61928.6562 - r_square: -5.0229 - val_loss: 51584.2891 - val_r_square: -5.0076\n",
      "Epoch 77/400\n",
      "8/8 [==============================] - 0s 25ms/step - loss: 57295.6523 - r_square: -4.5723 - val_loss: 53881.0234 - val_r_square: -5.2750\n",
      "Epoch 78/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 55316.0234 - r_square: -4.3798 - val_loss: 49814.2812 - val_r_square: -4.8014\n",
      "Epoch 79/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 53085.8086 - r_square: -4.1629 - val_loss: 50480.4805 - val_r_square: -4.8790\n",
      "Epoch 80/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 51722.0898 - r_square: -4.0303 - val_loss: 46962.4180 - val_r_square: -4.4693\n",
      "Epoch 81/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 50430.6992 - r_square: -3.9047 - val_loss: 47184.6211 - val_r_square: -4.4952\n",
      "Epoch 82/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 48461.7734 - r_square: -3.7132 - val_loss: 44680.4688 - val_r_square: -4.2035\n",
      "Epoch 83/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 46980.4023 - r_square: -3.5691 - val_loss: 43025.8359 - val_r_square: -4.0108\n",
      "Epoch 84/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 45532.1328 - r_square: -3.4282 - val_loss: 41571.3281 - val_r_square: -3.8414\n",
      "Epoch 85/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 43461.0234 - r_square: -3.2268 - val_loss: 43495.4336 - val_r_square: -4.0655\n",
      "Epoch 86/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 43384.8594 - r_square: -3.2194 - val_loss: 41509.1758 - val_r_square: -3.8342\n",
      "Epoch 87/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 42007.2109 - r_square: -3.0854 - val_loss: 39524.0117 - val_r_square: -3.6030\n",
      "Epoch 88/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 41277.3945 - r_square: -3.0144 - val_loss: 38390.5664 - val_r_square: -3.4710\n",
      "Epoch 89/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 40500.0469 - r_square: -2.9388 - val_loss: 37988.5781 - val_r_square: -3.4242\n",
      "Epoch 90/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 39453.2070 - r_square: -2.8370 - val_loss: 38969.7305 - val_r_square: -3.5385\n",
      "Epoch 91/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 38623.3203 - r_square: -2.7563 - val_loss: 37714.5938 - val_r_square: -3.3923\n",
      "Epoch 92/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 37354.8398 - r_square: -2.6330 - val_loss: 35789.0078 - val_r_square: -3.1680\n",
      "Epoch 93/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 36312.1953 - r_square: -2.5316 - val_loss: 34361.2500 - val_r_square: -3.0018\n",
      "Epoch 94/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 35419.7148 - r_square: -2.4447 - val_loss: 33634.9688 - val_r_square: -2.9172\n",
      "Epoch 95/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 34508.4219 - r_square: -2.3561 - val_loss: 33159.4375 - val_r_square: -2.8618\n",
      "Epoch 96/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 33895.6836 - r_square: -2.2965 - val_loss: 32484.5684 - val_r_square: -2.7832\n",
      "Epoch 97/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 33246.2812 - r_square: -2.2334 - val_loss: 31499.2461 - val_r_square: -2.6684\n",
      "Epoch 98/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 32663.3711 - r_square: -2.1767 - val_loss: 30723.0664 - val_r_square: -2.5780\n",
      "Epoch 99/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 32109.5430 - r_square: -2.1228 - val_loss: 30355.8848 - val_r_square: -2.5353\n",
      "Epoch 100/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 31578.2051 - r_square: -2.0712 - val_loss: 29945.9395 - val_r_square: -2.4875\n",
      "Epoch 101/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 31272.0859 - r_square: -2.0414 - val_loss: 29691.4863 - val_r_square: -2.4579\n",
      "Epoch 102/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 31020.0430 - r_square: -2.0168 - val_loss: 28986.9961 - val_r_square: -2.3759\n",
      "Epoch 103/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 30320.5547 - r_square: -1.9488 - val_loss: 28313.4785 - val_r_square: -2.2974\n",
      "Epoch 104/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 30042.5938 - r_square: -1.9218 - val_loss: 28469.0996 - val_r_square: -2.3155\n",
      "Epoch 105/400\n",
      "8/8 [==============================] - 0s 36ms/step - loss: 30256.1914 - r_square: -1.9426 - val_loss: 29016.6934 - val_r_square: -2.3793\n",
      "Epoch 106/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 29962.8301 - r_square: -1.9140 - val_loss: 30371.5684 - val_r_square: -2.5371\n",
      "Epoch 107/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 31171.6309 - r_square: -2.0316 - val_loss: 29376.4375 - val_r_square: -2.4212\n",
      "Epoch 108/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 30338.0312 - r_square: -1.9505 - val_loss: 27938.6152 - val_r_square: -2.2538\n",
      "Epoch 109/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 29479.7793 - r_square: -1.8671 - val_loss: 27572.6426 - val_r_square: -2.2111\n",
      "Epoch 110/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 28988.0605 - r_square: -1.8192 - val_loss: 27597.3574 - val_r_square: -2.2140\n",
      "Epoch 111/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 28166.4805 - r_square: -1.7393 - val_loss: 27143.2363 - val_r_square: -2.1611\n",
      "Epoch 112/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 27805.7559 - r_square: -1.7043 - val_loss: 26202.8008 - val_r_square: -2.0516\n",
      "Epoch 113/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 27276.7773 - r_square: -1.6528 - val_loss: 26369.0293 - val_r_square: -2.0710\n",
      "Epoch 114/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 26946.6074 - r_square: -1.6207 - val_loss: 26021.7266 - val_r_square: -2.0305\n",
      "Epoch 115/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 26585.0918 - r_square: -1.5855 - val_loss: 25424.4316 - val_r_square: -1.9610\n",
      "Epoch 116/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 26397.9277 - r_square: -1.5673 - val_loss: 24971.8516 - val_r_square: -1.9083\n",
      "Epoch 117/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 26329.5195 - r_square: -1.5607 - val_loss: 25457.5391 - val_r_square: -1.9648\n",
      "Epoch 118/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 26014.2344 - r_square: -1.5300 - val_loss: 24885.0859 - val_r_square: -1.8981\n",
      "Epoch 119/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 25536.6719 - r_square: -1.4836 - val_loss: 24513.9473 - val_r_square: -1.8549\n",
      "Epoch 120/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 25203.8926 - r_square: -1.4512 - val_loss: 24014.3223 - val_r_square: -1.7967\n",
      "Epoch 121/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 24956.5000 - r_square: -1.4272 - val_loss: 23625.1348 - val_r_square: -1.7514\n",
      "Epoch 122/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 24867.7344 - r_square: -1.4185 - val_loss: 22980.1914 - val_r_square: -1.6763\n",
      "Epoch 123/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 24147.5664 - r_square: -1.3485 - val_loss: 23450.3848 - val_r_square: -1.7311\n",
      "Epoch 124/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 24024.0625 - r_square: -1.3365 - val_loss: 22795.1504 - val_r_square: -1.6548\n",
      "Epoch 125/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 23707.1855 - r_square: -1.3057 - val_loss: 21589.7305 - val_r_square: -1.5144\n",
      "Epoch 126/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 23076.3223 - r_square: -1.2443 - val_loss: 21886.9023 - val_r_square: -1.5490\n",
      "Epoch 127/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 23191.7520 - r_square: -1.2555 - val_loss: 21105.2402 - val_r_square: -1.4579\n",
      "Epoch 128/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 23051.1523 - r_square: -1.2418 - val_loss: 20498.0430 - val_r_square: -1.3872\n",
      "Epoch 129/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 22881.3184 - r_square: -1.2253 - val_loss: 21541.0879 - val_r_square: -1.5087\n",
      "Epoch 130/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 22539.5449 - r_square: -1.1921 - val_loss: 20852.0938 - val_r_square: -1.4285\n",
      "Epoch 131/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 22446.3965 - r_square: -1.1830 - val_loss: 20521.2207 - val_r_square: -1.3899\n",
      "Epoch 132/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 21981.4082 - r_square: -1.1378 - val_loss: 20010.4746 - val_r_square: -1.3304\n",
      "Epoch 133/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 21603.0996 - r_square: -1.1010 - val_loss: 19291.1562 - val_r_square: -1.2467\n",
      "Epoch 134/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 21412.1621 - r_square: -1.0824 - val_loss: 19620.8887 - val_r_square: -1.2851\n",
      "Epoch 135/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 21088.8672 - r_square: -1.0510 - val_loss: 19639.7363 - val_r_square: -1.2873\n",
      "Epoch 136/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 20896.3340 - r_square: -1.0323 - val_loss: 19393.1895 - val_r_square: -1.2586\n",
      "Epoch 137/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 20611.6621 - r_square: -1.0046 - val_loss: 19099.7598 - val_r_square: -1.2244\n",
      "Epoch 138/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 20419.1797 - r_square: -0.9859 - val_loss: 18921.9688 - val_r_square: -1.2037\n",
      "Epoch 139/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 20231.9102 - r_square: -0.9677 - val_loss: 18682.4492 - val_r_square: -1.1758\n",
      "Epoch 140/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 20085.7148 - r_square: -0.9534 - val_loss: 18516.2695 - val_r_square: -1.1564\n",
      "Epoch 141/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 19709.6055 - r_square: -0.9169 - val_loss: 17816.1035 - val_r_square: -1.0749\n",
      "Epoch 142/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 19341.3691 - r_square: -0.8811 - val_loss: 17932.3770 - val_r_square: -1.0884\n",
      "Epoch 143/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 19202.2070 - r_square: -0.8675 - val_loss: 17677.5488 - val_r_square: -1.0587\n",
      "Epoch 144/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 18915.9629 - r_square: -0.8397 - val_loss: 18163.1973 - val_r_square: -1.1153\n",
      "Epoch 145/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 18620.5156 - r_square: -0.8109 - val_loss: 17809.5566 - val_r_square: -1.0741\n",
      "Epoch 146/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 18409.5996 - r_square: -0.7904 - val_loss: 17567.6094 - val_r_square: -1.0459\n",
      "Epoch 147/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 18466.3008 - r_square: -0.7959 - val_loss: 17304.7402 - val_r_square: -1.0153\n",
      "Epoch 148/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 18150.3262 - r_square: -0.7652 - val_loss: 17066.9844 - val_r_square: -0.9876\n",
      "Epoch 149/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17857.9004 - r_square: -0.7368 - val_loss: 16830.5898 - val_r_square: -0.9601\n",
      "Epoch 150/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 17758.2715 - r_square: -0.7271 - val_loss: 16882.0430 - val_r_square: -0.9661\n",
      "Epoch 151/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17554.9648 - r_square: -0.7073 - val_loss: 16755.6719 - val_r_square: -0.9514\n",
      "Epoch 152/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17404.2363 - r_square: -0.6927 - val_loss: 17154.4043 - val_r_square: -0.9978\n",
      "Epoch 153/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17266.7305 - r_square: -0.6793 - val_loss: 17341.1641 - val_r_square: -1.0196\n",
      "Epoch 154/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17421.2871 - r_square: -0.6943 - val_loss: 17190.0078 - val_r_square: -1.0020\n",
      "Epoch 155/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17153.3574 - r_square: -0.6683 - val_loss: 17167.1445 - val_r_square: -0.9993\n",
      "Epoch 156/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17216.8574 - r_square: -0.6744 - val_loss: 16941.9551 - val_r_square: -0.9731\n",
      "Epoch 157/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 16876.8281 - r_square: -0.6414 - val_loss: 16925.0312 - val_r_square: -0.9711\n",
      "Epoch 158/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17395.6621 - r_square: -0.6918 - val_loss: 17255.7207 - val_r_square: -1.0096\n",
      "Epoch 159/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17968.6094 - r_square: -0.7475 - val_loss: 16807.6992 - val_r_square: -0.9574\n",
      "Epoch 160/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17564.9531 - r_square: -0.7083 - val_loss: 17416.7109 - val_r_square: -1.0284\n",
      "Epoch 161/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17313.0488 - r_square: -0.6838 - val_loss: 16456.3340 - val_r_square: -0.9165\n",
      "Epoch 162/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17082.9746 - r_square: -0.6614 - val_loss: 17378.3809 - val_r_square: -1.0239\n",
      "Epoch 163/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16958.5078 - r_square: -0.6493 - val_loss: 17183.6523 - val_r_square: -1.0012\n",
      "Epoch 164/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 17195.7188 - r_square: -0.6724 - val_loss: 17170.1055 - val_r_square: -0.9997\n",
      "Epoch 165/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16791.8555 - r_square: -0.6331 - val_loss: 16642.6641 - val_r_square: -0.9382\n",
      "Epoch 166/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 16520.4746 - r_square: -0.6067 - val_loss: 16441.7207 - val_r_square: -0.9148\n",
      "Epoch 167/400\n",
      "8/8 [==============================] - 0s 30ms/step - loss: 16744.6133 - r_square: -0.6285 - val_loss: 16414.1973 - val_r_square: -0.9116\n",
      "Epoch 168/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 16641.9434 - r_square: -0.6185 - val_loss: 17121.2520 - val_r_square: -0.9940\n",
      "Epoch 169/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 17263.0566 - r_square: -0.6789 - val_loss: 16563.8770 - val_r_square: -0.9290\n",
      "Epoch 170/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16314.1504 - r_square: -0.5866 - val_loss: 16585.2520 - val_r_square: -0.9315\n",
      "Epoch 171/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16119.3076 - r_square: -0.5677 - val_loss: 16422.9980 - val_r_square: -0.9126\n",
      "Epoch 172/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 16006.6865 - r_square: -0.5567 - val_loss: 15872.5547 - val_r_square: -0.8485\n",
      "Epoch 173/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16014.2803 - r_square: -0.5575 - val_loss: 15970.0225 - val_r_square: -0.8599\n",
      "Epoch 174/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 15892.1953 - r_square: -0.5456 - val_loss: 16028.1270 - val_r_square: -0.8667\n",
      "Epoch 175/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 15887.3994 - r_square: -0.5451 - val_loss: 15824.6006 - val_r_square: -0.8430\n",
      "Epoch 176/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 16214.6963 - r_square: -0.5770 - val_loss: 15727.1836 - val_r_square: -0.8316\n",
      "Epoch 177/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16012.3545 - r_square: -0.5573 - val_loss: 16495.9824 - val_r_square: -0.9211\n",
      "Epoch 178/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 16004.9639 - r_square: -0.5566 - val_loss: 16300.5938 - val_r_square: -0.8984\n",
      "Epoch 179/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 16133.4775 - r_square: -0.5691 - val_loss: 16118.5215 - val_r_square: -0.8772\n",
      "Epoch 180/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 16270.1377 - r_square: -0.5823 - val_loss: 15778.8467 - val_r_square: -0.8376\n",
      "Epoch 181/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16069.1660 - r_square: -0.5628 - val_loss: 15490.9434 - val_r_square: -0.8041\n",
      "Epoch 182/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 16401.3086 - r_square: -0.5951 - val_loss: 15893.6553 - val_r_square: -0.8510\n",
      "Epoch 183/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16037.1758 - r_square: -0.5597 - val_loss: 15871.2070 - val_r_square: -0.8484\n",
      "Epoch 184/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15568.0537 - r_square: -0.5141 - val_loss: 15852.1562 - val_r_square: -0.8462\n",
      "Epoch 185/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15806.8232 - r_square: -0.5373 - val_loss: 15639.4961 - val_r_square: -0.8214\n",
      "Epoch 186/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 15402.7949 - r_square: -0.4980 - val_loss: 15536.9561 - val_r_square: -0.8095\n",
      "Epoch 187/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 15336.5713 - r_square: -0.4916 - val_loss: 15601.4668 - val_r_square: -0.8170\n",
      "Epoch 188/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 15363.1670 - r_square: -0.4942 - val_loss: 15554.5020 - val_r_square: -0.8115\n",
      "Epoch 189/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15184.6641 - r_square: -0.4768 - val_loss: 15637.0615 - val_r_square: -0.8211\n",
      "Epoch 190/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 15209.3857 - r_square: -0.4792 - val_loss: 15403.4199 - val_r_square: -0.7939\n",
      "Epoch 191/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 15391.9893 - r_square: -0.4969 - val_loss: 15665.6836 - val_r_square: -0.8244\n",
      "Epoch 192/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15039.0684 - r_square: -0.4626 - val_loss: 15602.6963 - val_r_square: -0.8171\n",
      "Epoch 193/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15726.1904 - r_square: -0.5294 - val_loss: 15881.0439 - val_r_square: -0.8495\n",
      "Epoch 194/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16197.4736 - r_square: -0.5753 - val_loss: 16324.7725 - val_r_square: -0.9012\n",
      "Epoch 195/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 15413.5352 - r_square: -0.4990 - val_loss: 15632.8271 - val_r_square: -0.8206\n",
      "Epoch 196/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 15332.3057 - r_square: -0.4911 - val_loss: 15231.7471 - val_r_square: -0.7739\n",
      "Epoch 197/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 15060.6836 - r_square: -0.4647 - val_loss: 15490.5576 - val_r_square: -0.8040\n",
      "Epoch 198/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 14930.9404 - r_square: -0.4521 - val_loss: 15761.6836 - val_r_square: -0.8356\n",
      "Epoch 199/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15227.9551 - r_square: -0.4810 - val_loss: 15757.4785 - val_r_square: -0.8351\n",
      "Epoch 200/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 14954.9990 - r_square: -0.4545 - val_loss: 15404.7715 - val_r_square: -0.7941\n",
      "Epoch 201/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14997.8252 - r_square: -0.4586 - val_loss: 15537.4971 - val_r_square: -0.8095\n",
      "Epoch 202/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14916.9404 - r_square: -0.4507 - val_loss: 15429.8711 - val_r_square: -0.7970\n",
      "Epoch 203/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 15073.6074 - r_square: -0.4660 - val_loss: 15558.1357 - val_r_square: -0.8119\n",
      "Epoch 204/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14904.7930 - r_square: -0.4496 - val_loss: 15484.3701 - val_r_square: -0.8033\n",
      "Epoch 205/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14771.7334 - r_square: -0.4366 - val_loss: 15274.3076 - val_r_square: -0.7789\n",
      "Epoch 206/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14845.4004 - r_square: -0.4438 - val_loss: 15185.3633 - val_r_square: -0.7685\n",
      "Epoch 207/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 14640.4512 - r_square: -0.4239 - val_loss: 15315.6553 - val_r_square: -0.7837\n",
      "Epoch 208/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14988.8594 - r_square: -0.4578 - val_loss: 15589.7559 - val_r_square: -0.8156\n",
      "Epoch 209/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 15087.6152 - r_square: -0.4673 - val_loss: 15742.9219 - val_r_square: -0.8334\n",
      "Epoch 210/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14592.5547 - r_square: -0.4192 - val_loss: 15443.5078 - val_r_square: -0.7986\n",
      "Epoch 211/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14510.3447 - r_square: -0.4112 - val_loss: 15570.4600 - val_r_square: -0.8134\n",
      "Epoch 212/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14995.2500 - r_square: -0.4584 - val_loss: 15605.9756 - val_r_square: -0.8175\n",
      "Epoch 213/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14587.6367 - r_square: -0.4187 - val_loss: 15498.0986 - val_r_square: -0.8049\n",
      "Epoch 214/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14425.8691 - r_square: -0.4030 - val_loss: 15187.2686 - val_r_square: -0.7687\n",
      "Epoch 215/400\n",
      "8/8 [==============================] - 0s 33ms/step - loss: 14429.3359 - r_square: -0.4033 - val_loss: 15314.5146 - val_r_square: -0.7835\n",
      "Epoch 216/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14598.9072 - r_square: -0.4198 - val_loss: 15216.2158 - val_r_square: -0.7721\n",
      "Epoch 217/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14619.6787 - r_square: -0.4218 - val_loss: 15076.2051 - val_r_square: -0.7558\n",
      "Epoch 218/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14459.3047 - r_square: -0.4063 - val_loss: 14949.2197 - val_r_square: -0.7410\n",
      "Epoch 219/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15083.6641 - r_square: -0.4670 - val_loss: 15818.0020 - val_r_square: -0.8422\n",
      "Epoch 220/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14656.5361 - r_square: -0.4254 - val_loss: 15871.3232 - val_r_square: -0.8484\n",
      "Epoch 221/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 14797.7461 - r_square: -0.4392 - val_loss: 14681.4600 - val_r_square: -0.7098\n",
      "Epoch 222/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15053.2998 - r_square: -0.4640 - val_loss: 14754.0283 - val_r_square: -0.7183\n",
      "Epoch 223/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14985.2607 - r_square: -0.4574 - val_loss: 14944.8135 - val_r_square: -0.7405\n",
      "Epoch 224/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14912.7520 - r_square: -0.4503 - val_loss: 15788.1221 - val_r_square: -0.8387\n",
      "Epoch 225/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14456.2559 - r_square: -0.4059 - val_loss: 16003.4980 - val_r_square: -0.8638\n",
      "Epoch 226/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 14530.5771 - r_square: -0.4132 - val_loss: 15453.9326 - val_r_square: -0.7998\n",
      "Epoch 227/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 14316.2861 - r_square: -0.3923 - val_loss: 14478.6016 - val_r_square: -0.6862\n",
      "Epoch 228/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 14139.5762 - r_square: -0.3752 - val_loss: 15601.1328 - val_r_square: -0.8169\n",
      "Epoch 229/400\n",
      "8/8 [==============================] - 0s 22ms/step - loss: 14621.6846 - r_square: -0.4221 - val_loss: 14822.5176 - val_r_square: -0.7262\n",
      "Epoch 230/400\n",
      "8/8 [==============================] - 0s 22ms/step - loss: 14176.2666 - r_square: -0.3787 - val_loss: 14772.5850 - val_r_square: -0.7204\n",
      "Epoch 231/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 14286.4619 - r_square: -0.3894 - val_loss: 14445.5039 - val_r_square: -0.6823\n",
      "Epoch 232/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14084.0830 - r_square: -0.3698 - val_loss: 14137.2070 - val_r_square: -0.6464\n",
      "Epoch 233/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14240.3799 - r_square: -0.3850 - val_loss: 14361.4775 - val_r_square: -0.6726\n",
      "Epoch 234/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14131.5479 - r_square: -0.3744 - val_loss: 14299.8887 - val_r_square: -0.6654\n",
      "Epoch 235/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 14087.3516 - r_square: -0.3701 - val_loss: 14435.7617 - val_r_square: -0.6812\n",
      "Epoch 236/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14057.2168 - r_square: -0.3671 - val_loss: 14030.2100 - val_r_square: -0.6340\n",
      "Epoch 237/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14107.5918 - r_square: -0.3720 - val_loss: 14976.2393 - val_r_square: -0.7442\n",
      "Epoch 238/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14628.5312 - r_square: -0.4227 - val_loss: 14096.6660 - val_r_square: -0.6417\n",
      "Epoch 239/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14475.9600 - r_square: -0.4079 - val_loss: 14529.5488 - val_r_square: -0.6921\n",
      "Epoch 240/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 14213.8438 - r_square: -0.3824 - val_loss: 14445.5664 - val_r_square: -0.6823\n",
      "Epoch 241/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14414.2236 - r_square: -0.4019 - val_loss: 14359.5615 - val_r_square: -0.6723\n",
      "Epoch 242/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14191.0986 - r_square: -0.3802 - val_loss: 14170.4062 - val_r_square: -0.6503\n",
      "Epoch 243/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14000.2305 - r_square: -0.3616 - val_loss: 15501.6533 - val_r_square: -0.8053\n",
      "Epoch 244/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14757.2314 - r_square: -0.4352 - val_loss: 14392.7490 - val_r_square: -0.6762\n",
      "Epoch 245/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14657.2246 - r_square: -0.4255 - val_loss: 15114.7256 - val_r_square: -0.7603\n",
      "Epoch 246/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14894.5371 - r_square: -0.4486 - val_loss: 16044.0742 - val_r_square: -0.8685\n",
      "Epoch 247/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14876.6729 - r_square: -0.4469 - val_loss: 15384.4453 - val_r_square: -0.7917\n",
      "Epoch 248/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14502.9502 - r_square: -0.4105 - val_loss: 15399.2617 - val_r_square: -0.7934\n",
      "Epoch 249/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14914.7812 - r_square: -0.4505 - val_loss: 14803.7393 - val_r_square: -0.7241\n",
      "Epoch 250/400\n",
      "8/8 [==============================] - 0s 34ms/step - loss: 14502.2314 - r_square: -0.4104 - val_loss: 14981.3789 - val_r_square: -0.7447\n",
      "Epoch 251/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14529.0889 - r_square: -0.4130 - val_loss: 14399.3818 - val_r_square: -0.6770\n",
      "Epoch 252/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14516.2773 - r_square: -0.4118 - val_loss: 14778.6016 - val_r_square: -0.7211\n",
      "Epoch 253/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14288.0527 - r_square: -0.3896 - val_loss: 14161.3232 - val_r_square: -0.6492\n",
      "Epoch 254/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14519.6025 - r_square: -0.4121 - val_loss: 14319.5479 - val_r_square: -0.6677\n",
      "Epoch 255/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 13966.2666 - r_square: -0.3583 - val_loss: 14241.6689 - val_r_square: -0.6586\n",
      "Epoch 256/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13818.5713 - r_square: -0.3439 - val_loss: 14613.2578 - val_r_square: -0.7019\n",
      "Epoch 257/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13680.0371 - r_square: -0.3304 - val_loss: 15155.1367 - val_r_square: -0.7650\n",
      "Epoch 258/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 14045.8643 - r_square: -0.3660 - val_loss: 14687.0459 - val_r_square: -0.7105\n",
      "Epoch 259/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14892.7471 - r_square: -0.4484 - val_loss: 14430.7988 - val_r_square: -0.6806\n",
      "Epoch 260/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14366.1582 - r_square: -0.3972 - val_loss: 14871.0762 - val_r_square: -0.7319\n",
      "Epoch 261/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 14237.6025 - r_square: -0.3847 - val_loss: 14657.2031 - val_r_square: -0.7070\n",
      "Epoch 262/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 14226.1611 - r_square: -0.3836 - val_loss: 15008.2412 - val_r_square: -0.7479\n",
      "Epoch 263/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13905.1387 - r_square: -0.3523 - val_loss: 14488.4238 - val_r_square: -0.6873\n",
      "Epoch 264/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14071.0576 - r_square: -0.3685 - val_loss: 14342.7402 - val_r_square: -0.6704\n",
      "Epoch 265/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13694.0176 - r_square: -0.3318 - val_loss: 13866.8867 - val_r_square: -0.6150\n",
      "Epoch 266/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13520.7988 - r_square: -0.3150 - val_loss: 14211.9365 - val_r_square: -0.6551\n",
      "Epoch 267/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 13468.6943 - r_square: -0.3099 - val_loss: 13805.6533 - val_r_square: -0.6078\n",
      "Epoch 268/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13604.2793 - r_square: -0.3231 - val_loss: 15037.5352 - val_r_square: -0.7513\n",
      "Epoch 269/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13784.9160 - r_square: -0.3407 - val_loss: 13807.7109 - val_r_square: -0.6081\n",
      "Epoch 270/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 13683.1318 - r_square: -0.3308 - val_loss: 15158.4521 - val_r_square: -0.7654\n",
      "Epoch 271/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14105.6191 - r_square: -0.3719 - val_loss: 14371.4404 - val_r_square: -0.6737\n",
      "Epoch 272/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13632.9111 - r_square: -0.3259 - val_loss: 14952.3311 - val_r_square: -0.7414\n",
      "Epoch 273/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13459.2090 - r_square: -0.3090 - val_loss: 14435.1748 - val_r_square: -0.6811\n",
      "Epoch 274/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13476.9883 - r_square: -0.3107 - val_loss: 14105.4150 - val_r_square: -0.6427\n",
      "Epoch 275/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13298.5752 - r_square: -0.2934 - val_loss: 14065.9521 - val_r_square: -0.6381\n",
      "Epoch 276/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13327.9639 - r_square: -0.2962 - val_loss: 13923.7139 - val_r_square: -0.6216\n",
      "Epoch 277/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13519.4238 - r_square: -0.3148 - val_loss: 14586.3730 - val_r_square: -0.6987\n",
      "Epoch 278/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13322.4531 - r_square: -0.2957 - val_loss: 13866.6836 - val_r_square: -0.6149\n",
      "Epoch 279/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13780.5293 - r_square: -0.3402 - val_loss: 14050.3330 - val_r_square: -0.6363\n",
      "Epoch 280/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13172.5156 - r_square: -0.2811 - val_loss: 13709.8018 - val_r_square: -0.5967\n",
      "Epoch 281/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13132.5303 - r_square: -0.2772 - val_loss: 13227.9883 - val_r_square: -0.5405\n",
      "Epoch 282/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 13181.9551 - r_square: -0.2820 - val_loss: 13888.5176 - val_r_square: -0.6175\n",
      "Epoch 283/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14399.3945 - r_square: -0.4004 - val_loss: 14368.3818 - val_r_square: -0.6734\n",
      "Epoch 284/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13578.8730 - r_square: -0.3206 - val_loss: 15764.5225 - val_r_square: -0.8360\n",
      "Epoch 285/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 18078.1172 - r_square: -0.7582 - val_loss: 14746.3447 - val_r_square: -0.7174\n",
      "Epoch 286/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 15328.4248 - r_square: -0.4908 - val_loss: 15006.0215 - val_r_square: -0.7476\n",
      "Epoch 287/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14474.2715 - r_square: -0.4077 - val_loss: 15000.4082 - val_r_square: -0.7470\n",
      "Epoch 288/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14135.8281 - r_square: -0.3748 - val_loss: 15340.6484 - val_r_square: -0.7866\n",
      "Epoch 289/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13365.2412 - r_square: -0.2998 - val_loss: 14498.4258 - val_r_square: -0.6885\n",
      "Epoch 290/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13192.0000 - r_square: -0.2830 - val_loss: 13752.6689 - val_r_square: -0.6017\n",
      "Epoch 291/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13333.6846 - r_square: -0.2968 - val_loss: 13943.5957 - val_r_square: -0.6239\n",
      "Epoch 292/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 13269.3184 - r_square: -0.2905 - val_loss: 14048.6592 - val_r_square: -0.6361\n",
      "Epoch 293/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13213.9160 - r_square: -0.2851 - val_loss: 14002.9434 - val_r_square: -0.6308\n",
      "Epoch 294/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13498.0469 - r_square: -0.3127 - val_loss: 15059.8232 - val_r_square: -0.7539\n",
      "Epoch 295/400\n",
      "8/8 [==============================] - 0s 35ms/step - loss: 13357.2529 - r_square: -0.2991 - val_loss: 13733.4434 - val_r_square: -0.5994\n",
      "Epoch 296/400\n",
      "8/8 [==============================] - 0s 15ms/step - loss: 13433.0312 - r_square: -0.3064 - val_loss: 15264.9639 - val_r_square: -0.7778\n",
      "Epoch 297/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13310.2236 - r_square: -0.2945 - val_loss: 14090.1221 - val_r_square: -0.6410\n",
      "Epoch 298/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13078.3447 - r_square: -0.2719 - val_loss: 13928.2793 - val_r_square: -0.6221\n",
      "Epoch 299/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12908.3066 - r_square: -0.2554 - val_loss: 13735.9346 - val_r_square: -0.5997\n",
      "Epoch 300/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12671.6104 - r_square: -0.2324 - val_loss: 13736.3994 - val_r_square: -0.5998\n",
      "Epoch 301/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13553.6006 - r_square: -0.3182 - val_loss: 14505.2842 - val_r_square: -0.6893\n",
      "Epoch 302/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13333.1396 - r_square: -0.2967 - val_loss: 13483.8164 - val_r_square: -0.5703\n",
      "Epoch 303/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 12945.1367 - r_square: -0.2590 - val_loss: 14346.7490 - val_r_square: -0.6708\n",
      "Epoch 304/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13487.6143 - r_square: -0.3117 - val_loss: 12603.4756 - val_r_square: -0.4678\n",
      "Epoch 305/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13505.5645 - r_square: -0.3135 - val_loss: 13613.9541 - val_r_square: -0.5855\n",
      "Epoch 306/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12985.2256 - r_square: -0.2629 - val_loss: 13189.5479 - val_r_square: -0.5361\n",
      "Epoch 307/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12904.3887 - r_square: -0.2550 - val_loss: 13431.7969 - val_r_square: -0.5643\n",
      "Epoch 308/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12722.1270 - r_square: -0.2373 - val_loss: 13337.4824 - val_r_square: -0.5533\n",
      "Epoch 309/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13134.2471 - r_square: -0.2774 - val_loss: 14260.2510 - val_r_square: -0.6608\n",
      "Epoch 310/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 20004.7207 - r_square: -0.9456 - val_loss: 15445.6309 - val_r_square: -0.7988\n",
      "Epoch 311/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12996.3018 - r_square: -0.2640 - val_loss: 13511.0762 - val_r_square: -0.5735\n",
      "Epoch 312/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12933.2500 - r_square: -0.2578 - val_loss: 12885.9805 - val_r_square: -0.5007\n",
      "Epoch 313/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13636.3203 - r_square: -0.3262 - val_loss: 14115.1104 - val_r_square: -0.6439\n",
      "Epoch 314/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13156.4443 - r_square: -0.2795 - val_loss: 15170.1416 - val_r_square: -0.7667\n",
      "Epoch 315/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13037.8096 - r_square: -0.2680 - val_loss: 13320.5010 - val_r_square: -0.5513\n",
      "Epoch 316/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 12765.2754 - r_square: -0.2415 - val_loss: 14092.7754 - val_r_square: -0.6413\n",
      "Epoch 317/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12478.7754 - r_square: -0.2136 - val_loss: 12812.2793 - val_r_square: -0.4921\n",
      "Epoch 318/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12792.9834 - r_square: -0.2442 - val_loss: 14418.6484 - val_r_square: -0.6792\n",
      "Epoch 319/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12613.8945 - r_square: -0.2268 - val_loss: 13573.5391 - val_r_square: -0.5808\n",
      "Epoch 320/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12310.9365 - r_square: -0.1973 - val_loss: 13622.5146 - val_r_square: -0.5865\n",
      "Epoch 321/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12298.1035 - r_square: -0.1961 - val_loss: 13666.4922 - val_r_square: -0.5916\n",
      "Epoch 322/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12224.8066 - r_square: -0.1889 - val_loss: 13784.6650 - val_r_square: -0.6054\n",
      "Epoch 323/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12485.8545 - r_square: -0.2143 - val_loss: 13482.6152 - val_r_square: -0.5702\n",
      "Epoch 324/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 12213.7480 - r_square: -0.1879 - val_loss: 13392.0723 - val_r_square: -0.5597\n",
      "Epoch 325/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12058.5283 - r_square: -0.1727 - val_loss: 14211.5244 - val_r_square: -0.6551\n",
      "Epoch 326/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12139.6367 - r_square: -0.1807 - val_loss: 12942.2441 - val_r_square: -0.5073\n",
      "Epoch 327/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12444.7480 - r_square: -0.2103 - val_loss: 13716.0664 - val_r_square: -0.5974\n",
      "Epoch 328/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12500.0762 - r_square: -0.2157 - val_loss: 13723.8857 - val_r_square: -0.5983\n",
      "Epoch 329/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13376.7891 - r_square: -0.3010 - val_loss: 13778.0439 - val_r_square: -0.6046\n",
      "Epoch 330/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12642.0625 - r_square: -0.2295 - val_loss: 14324.9062 - val_r_square: -0.6683\n",
      "Epoch 331/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12532.1680 - r_square: -0.2188 - val_loss: 13262.2949 - val_r_square: -0.5445\n",
      "Epoch 332/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 13104.7334 - r_square: -0.2745 - val_loss: 12841.9980 - val_r_square: -0.4956\n",
      "Epoch 333/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13157.7832 - r_square: -0.2797 - val_loss: 14318.8467 - val_r_square: -0.6676\n",
      "Epoch 334/400\n",
      "8/8 [==============================] - 0s 28ms/step - loss: 14197.6943 - r_square: -0.3808 - val_loss: 13659.1318 - val_r_square: -0.5908\n",
      "Epoch 335/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 14155.6074 - r_square: -0.3767 - val_loss: 17601.4570 - val_r_square: -1.0499\n",
      "Epoch 336/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 25611.5723 - r_square: -1.4909 - val_loss: 30790.1270 - val_r_square: -2.5859\n",
      "Epoch 337/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 25633.6797 - r_square: -1.4930 - val_loss: 17844.7949 - val_r_square: -1.0782\n",
      "Epoch 338/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 21465.4238 - r_square: -1.0876 - val_loss: 24733.7461 - val_r_square: -1.8805\n",
      "Epoch 339/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 18231.7988 - r_square: -0.7731 - val_loss: 16641.9707 - val_r_square: -0.9381\n",
      "Epoch 340/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 17884.9434 - r_square: -0.7394 - val_loss: 19175.7285 - val_r_square: -1.2332\n",
      "Epoch 341/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 17602.6348 - r_square: -0.7120 - val_loss: 23863.9883 - val_r_square: -1.7792\n",
      "Epoch 342/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 18878.5000 - r_square: -0.8360 - val_loss: 20223.0684 - val_r_square: -1.3552\n",
      "Epoch 343/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16500.9395 - r_square: -0.6048 - val_loss: 18751.7461 - val_r_square: -1.1839\n",
      "Epoch 344/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15353.1777 - r_square: -0.4932 - val_loss: 15675.2344 - val_r_square: -0.8256\n",
      "Epoch 345/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14651.2188 - r_square: -0.4249 - val_loss: 15142.3760 - val_r_square: -0.7635\n",
      "Epoch 346/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14836.4883 - r_square: -0.4429 - val_loss: 22048.4707 - val_r_square: -1.5678\n",
      "Epoch 347/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 18978.7754 - r_square: -0.8458 - val_loss: 19328.4316 - val_r_square: -1.2510\n",
      "Epoch 348/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 15761.6299 - r_square: -0.5329 - val_loss: 18551.1973 - val_r_square: -1.1605\n",
      "Epoch 349/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14170.9443 - r_square: -0.3782 - val_loss: 14723.7246 - val_r_square: -0.7147\n",
      "Epoch 350/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14084.6240 - r_square: -0.3698 - val_loss: 15485.7773 - val_r_square: -0.8035\n",
      "Epoch 351/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 14461.4805 - r_square: -0.4065 - val_loss: 14552.2051 - val_r_square: -0.6948\n",
      "Epoch 352/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13530.6611 - r_square: -0.3159 - val_loss: 14593.1299 - val_r_square: -0.6995\n",
      "Epoch 353/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 13355.9424 - r_square: -0.2989 - val_loss: 15514.6689 - val_r_square: -0.8069\n",
      "Epoch 354/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13148.6123 - r_square: -0.2788 - val_loss: 14251.3320 - val_r_square: -0.6597\n",
      "Epoch 355/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 13211.4287 - r_square: -0.2849 - val_loss: 16759.7812 - val_r_square: -0.9519\n",
      "Epoch 356/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13734.6221 - r_square: -0.3358 - val_loss: 15758.7793 - val_r_square: -0.8353\n",
      "Epoch 357/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12941.6406 - r_square: -0.2586 - val_loss: 15628.1182 - val_r_square: -0.8201\n",
      "Epoch 358/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12947.5850 - r_square: -0.2592 - val_loss: 15237.7812 - val_r_square: -0.7746\n",
      "Epoch 359/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 12692.3389 - r_square: -0.2344 - val_loss: 14581.0742 - val_r_square: -0.6981\n",
      "Epoch 360/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12548.7139 - r_square: -0.2204 - val_loss: 14241.0439 - val_r_square: -0.6585\n",
      "Epoch 361/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13295.4180 - r_square: -0.2931 - val_loss: 14673.5283 - val_r_square: -0.7089\n",
      "Epoch 362/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 13725.4365 - r_square: -0.3349 - val_loss: 14025.6514 - val_r_square: -0.6334\n",
      "Epoch 363/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13642.2803 - r_square: -0.3268 - val_loss: 16667.9336 - val_r_square: -0.9412\n",
      "Epoch 364/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 14852.0498 - r_square: -0.4444 - val_loss: 14634.8184 - val_r_square: -0.7044\n",
      "Epoch 365/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13021.6484 - r_square: -0.2664 - val_loss: 13561.1182 - val_r_square: -0.5793\n",
      "Epoch 366/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12666.1797 - r_square: -0.2319 - val_loss: 16170.1211 - val_r_square: -0.8832\n",
      "Epoch 367/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13087.0576 - r_square: -0.2728 - val_loss: 13369.0801 - val_r_square: -0.5570\n",
      "Epoch 368/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 13053.3174 - r_square: -0.2695 - val_loss: 14077.1475 - val_r_square: -0.6394\n",
      "Epoch 369/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12511.3887 - r_square: -0.2168 - val_loss: 14019.5410 - val_r_square: -0.6327\n",
      "Epoch 370/400\n",
      "8/8 [==============================] - 0s 32ms/step - loss: 12411.4033 - r_square: -0.2071 - val_loss: 14586.7725 - val_r_square: -0.6988\n",
      "Epoch 371/400\n",
      "8/8 [==============================] - 0s 20ms/step - loss: 12462.2471 - r_square: -0.2120 - val_loss: 13644.8496 - val_r_square: -0.5891\n",
      "Epoch 372/400\n",
      "8/8 [==============================] - 0s 21ms/step - loss: 12331.9414 - r_square: -0.1993 - val_loss: 13370.4805 - val_r_square: -0.5571\n",
      "Epoch 373/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12635.4189 - r_square: -0.2289 - val_loss: 14853.2617 - val_r_square: -0.7298\n",
      "Epoch 374/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12767.3350 - r_square: -0.2417 - val_loss: 13221.5020 - val_r_square: -0.5398\n",
      "Epoch 375/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 12833.1348 - r_square: -0.2481 - val_loss: 15704.3887 - val_r_square: -0.8290\n",
      "Epoch 376/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12741.9619 - r_square: -0.2392 - val_loss: 12765.8516 - val_r_square: -0.4867\n",
      "Epoch 377/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 12967.0225 - r_square: -0.2611 - val_loss: 20142.1348 - val_r_square: -1.3458\n",
      "Epoch 378/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 16950.7031 - r_square: -0.6486 - val_loss: 15681.3887 - val_r_square: -0.8263\n",
      "Epoch 379/400\n",
      "8/8 [==============================] - 0s 19ms/step - loss: 14411.5908 - r_square: -0.4016 - val_loss: 14741.0244 - val_r_square: -0.7168\n",
      "Epoch 380/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 13501.2168 - r_square: -0.3131 - val_loss: 13760.4541 - val_r_square: -0.6026\n",
      "Epoch 381/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12929.9609 - r_square: -0.2575 - val_loss: 12818.5898 - val_r_square: -0.4929\n",
      "Epoch 382/400\n",
      "8/8 [==============================] - 0s 16ms/step - loss: 12427.8223 - r_square: -0.2087 - val_loss: 13420.3877 - val_r_square: -0.5630\n",
      "Epoch 383/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12713.3672 - r_square: -0.2364 - val_loss: 13823.0303 - val_r_square: -0.6098\n",
      "Epoch 384/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13124.4424 - r_square: -0.2764 - val_loss: 12940.6025 - val_r_square: -0.5071\n",
      "Epoch 385/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12653.8945 - r_square: -0.2307 - val_loss: 13613.4219 - val_r_square: -0.5854\n",
      "Epoch 386/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12556.2666 - r_square: -0.2212 - val_loss: 12643.1934 - val_r_square: -0.4724\n",
      "Epoch 387/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12345.1953 - r_square: -0.2006 - val_loss: 13492.2100 - val_r_square: -0.5713\n",
      "Epoch 388/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12398.1416 - r_square: -0.2058 - val_loss: 13289.2256 - val_r_square: -0.5477\n",
      "Epoch 389/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12095.0312 - r_square: -0.1763 - val_loss: 12984.3779 - val_r_square: -0.5122\n",
      "Epoch 390/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12028.5547 - r_square: -0.1698 - val_loss: 14068.7803 - val_r_square: -0.6385\n",
      "Epoch 391/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12359.9160 - r_square: -0.2021 - val_loss: 12485.3037 - val_r_square: -0.4541\n",
      "Epoch 392/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12858.4082 - r_square: -0.2506 - val_loss: 14712.5195 - val_r_square: -0.7134\n",
      "Epoch 393/400\n",
      "8/8 [==============================] - 0s 18ms/step - loss: 12143.6035 - r_square: -0.1810 - val_loss: 13990.9668 - val_r_square: -0.6294\n",
      "Epoch 394/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12029.8584 - r_square: -0.1700 - val_loss: 13300.0938 - val_r_square: -0.5489\n",
      "Epoch 395/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12308.5439 - r_square: -0.1971 - val_loss: 15762.6924 - val_r_square: -0.8357\n",
      "Epoch 396/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12486.3096 - r_square: -0.2144 - val_loss: 12925.0117 - val_r_square: -0.5053\n",
      "Epoch 397/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 14028.6387 - r_square: -0.3644 - val_loss: 17799.0430 - val_r_square: -1.0729\n",
      "Epoch 398/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 13435.7334 - r_square: -0.3067 - val_loss: 13764.1963 - val_r_square: -0.6030\n",
      "Epoch 399/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 12431.4033 - r_square: -0.2090 - val_loss: 13223.1660 - val_r_square: -0.5400\n",
      "Epoch 400/400\n",
      "8/8 [==============================] - 0s 17ms/step - loss: 11907.5742 - r_square: -0.1581 - val_loss: 12312.1816 - val_r_square: -0.4339\n"
     ]
    }
   ],
   "source": [
    "#Another Type of RNN similar to LSTM\n",
    "\n",
    "from tensorflow.keras.layers import GRU, Dense\n",
    "\n",
    "model_gru = Sequential()\n",
    "model_gru.add(GRU(64, activation='relu', input_shape=(10, 5)))\n",
    "model_gru.add(Dense(1, activation='linear'))\n",
    "\n",
    "model_gru.compile(loss='mean_squared_error', optimizer=tf.keras.optimizers.Adam(learning_rate=0.1), metrics=[tfa.metrics.RSquare()])\n",
    "\n",
    "\n",
    "history_gru = model_gru.fit(X_train_rs, y_train1, epochs=400, batch_size=128, validation_data=(X_val_rs, y_val1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 384,
   "id": "31449999-df07-4a4a-ac42-22f16a189494",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 0s 3ms/step\n",
      "108.20584339781315\n"
     ]
    }
   ],
   "source": [
    "y_pred_gru_tr = model_gru.predict(X_train_rs)\n",
    "gru_rmse_train = np.sqrt(mean_squared_error(y_train1, y_pred_gru_tr))\n",
    "print(gru_rmse_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 550,
   "id": "3ba22c26-9222-4127-a708-562be82072c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "## ------------- GRU with more layers, Batch Normalization and Dropout -------------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "id": "775eda2b-f8c3-4580-83f4-fa0ec9b65d9b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/400\n",
      "8/8 [==============================] - 14s 528ms/step - loss: 393061.7500 - r_square: -35.7987 - val_loss: 348800.1250 - val_r_square: -42.6762\n",
      "Epoch 2/400\n",
      "8/8 [==============================] - 1s 71ms/step - loss: 350561.5312 - r_square: -31.8199 - val_loss: 270866.5000 - val_r_square: -32.9174\n",
      "Epoch 3/400\n",
      "8/8 [==============================] - 1s 63ms/step - loss: 259056.7812 - r_square: -23.2532 - val_loss: 184746.2812 - val_r_square: -22.1336\n",
      "Epoch 4/400\n",
      "8/8 [==============================] - 0s 56ms/step - loss: 139109.3125 - r_square: -12.0235 - val_loss: 59098.7227 - val_r_square: -6.4002\n",
      "Epoch 5/400\n",
      "8/8 [==============================] - 1s 70ms/step - loss: 39230.9180 - r_square: -2.6728 - val_loss: 6619.4355 - val_r_square: 0.1711\n",
      "Epoch 6/400\n",
      "8/8 [==============================] - 0s 60ms/step - loss: 14189.5283 - r_square: -0.3284 - val_loss: 22029.9004 - val_r_square: -1.7585\n",
      "Epoch 7/400\n",
      "8/8 [==============================] - 0s 54ms/step - loss: 20625.5723 - r_square: -0.9310 - val_loss: 15959.2119 - val_r_square: -0.9984\n",
      "Epoch 8/400\n",
      "8/8 [==============================] - 1s 78ms/step - loss: 13163.1680 - r_square: -0.2324 - val_loss: 8167.8760 - val_r_square: -0.0228\n",
      "Epoch 9/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 12481.8457 - r_square: -0.1686 - val_loss: 7172.5161 - val_r_square: 0.1019\n",
      "Epoch 10/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12483.2529 - r_square: -0.1687 - val_loss: 6837.6548 - val_r_square: 0.1438\n",
      "Epoch 11/400\n",
      "8/8 [==============================] - 0s 53ms/step - loss: 11058.4434 - r_square: -0.0353 - val_loss: 6802.0894 - val_r_square: 0.1483\n",
      "Epoch 12/400\n",
      "8/8 [==============================] - 0s 56ms/step - loss: 11668.2988 - r_square: -0.0924 - val_loss: 7545.1934 - val_r_square: 0.0552\n",
      "Epoch 13/400\n",
      "8/8 [==============================] - 0s 58ms/step - loss: 11482.1279 - r_square: -0.0750 - val_loss: 6711.3672 - val_r_square: 0.1596\n",
      "Epoch 14/400\n",
      "8/8 [==============================] - 1s 68ms/step - loss: 11866.2930 - r_square: -0.1110 - val_loss: 7340.5083 - val_r_square: 0.0808\n",
      "Epoch 15/400\n",
      "8/8 [==============================] - 0s 60ms/step - loss: 10993.0430 - r_square: -0.0292 - val_loss: 8842.5488 - val_r_square: -0.1072\n",
      "Epoch 16/400\n",
      "8/8 [==============================] - 0s 58ms/step - loss: 11534.1045 - r_square: -0.0798 - val_loss: 6518.0391 - val_r_square: 0.1838\n",
      "Epoch 17/400\n",
      "8/8 [==============================] - 0s 59ms/step - loss: 10780.0293 - r_square: -0.0092 - val_loss: 8581.3965 - val_r_square: -0.0745\n",
      "Epoch 18/400\n",
      "8/8 [==============================] - 0s 57ms/step - loss: 11118.3154 - r_square: -0.0409 - val_loss: 7213.1782 - val_r_square: 0.0968\n",
      "Epoch 19/400\n",
      "8/8 [==============================] - 1s 66ms/step - loss: 11462.6611 - r_square: -0.0731 - val_loss: 7121.6411 - val_r_square: 0.1082\n",
      "Epoch 20/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11192.5244 - r_square: -0.0479 - val_loss: 7252.1152 - val_r_square: 0.0919\n",
      "Epoch 21/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 10940.5645 - r_square: -0.0243 - val_loss: 8143.2710 - val_r_square: -0.0197\n",
      "Epoch 22/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 10945.3135 - r_square: -0.0247 - val_loss: 6459.0400 - val_r_square: 0.1912\n",
      "Epoch 23/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 10654.7705 - r_square: 0.0025 - val_loss: 6105.2588 - val_r_square: 0.2355\n",
      "Epoch 24/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10654.7080 - r_square: 0.0025 - val_loss: 6367.3164 - val_r_square: 0.2027\n",
      "Epoch 25/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 10849.9688 - r_square: -0.0158 - val_loss: 8087.8228 - val_r_square: -0.0127\n",
      "Epoch 26/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11204.9248 - r_square: -0.0490 - val_loss: 6819.6606 - val_r_square: 0.1461\n",
      "Epoch 27/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11598.5195 - r_square: -0.0859 - val_loss: 6246.6074 - val_r_square: 0.2178\n",
      "Epoch 28/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11639.0029 - r_square: -0.0897 - val_loss: 6166.7314 - val_r_square: 0.2278\n",
      "Epoch 29/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11238.0635 - r_square: -0.0521 - val_loss: 5878.0371 - val_r_square: 0.2640\n",
      "Epoch 30/400\n",
      "8/8 [==============================] - 1s 70ms/step - loss: 10917.0312 - r_square: -0.0221 - val_loss: 5540.4185 - val_r_square: 0.3062\n",
      "Epoch 31/400\n",
      "8/8 [==============================] - 0s 54ms/step - loss: 10309.1875 - r_square: 0.0348 - val_loss: 5577.6841 - val_r_square: 0.3016\n",
      "Epoch 32/400\n",
      "8/8 [==============================] - 0s 64ms/step - loss: 10633.7900 - r_square: 0.0045 - val_loss: 5569.5405 - val_r_square: 0.3026\n",
      "Epoch 33/400\n",
      "8/8 [==============================] - 1s 66ms/step - loss: 11402.9521 - r_square: -0.0676 - val_loss: 5825.7988 - val_r_square: 0.2705\n",
      "Epoch 34/400\n",
      "8/8 [==============================] - 1s 74ms/step - loss: 10918.9961 - r_square: -0.0222 - val_loss: 5637.0659 - val_r_square: 0.2941\n",
      "Epoch 35/400\n",
      "8/8 [==============================] - 0s 57ms/step - loss: 11643.2275 - r_square: -0.0901 - val_loss: 6274.5332 - val_r_square: 0.2143\n",
      "Epoch 36/400\n",
      "8/8 [==============================] - 1s 66ms/step - loss: 11631.5820 - r_square: -0.0890 - val_loss: 6564.8672 - val_r_square: 0.1780\n",
      "Epoch 37/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11976.7383 - r_square: -0.1213 - val_loss: 10452.8936 - val_r_square: -0.3089\n",
      "Epoch 38/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11912.0928 - r_square: -0.1152 - val_loss: 8344.6738 - val_r_square: -0.0449\n",
      "Epoch 39/400\n",
      "8/8 [==============================] - 0s 55ms/step - loss: 11556.2539 - r_square: -0.0819 - val_loss: 7763.8359 - val_r_square: 0.0278\n",
      "Epoch 40/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 12277.7100 - r_square: -0.1495 - val_loss: 6683.0200 - val_r_square: 0.1632\n",
      "Epoch 41/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11375.4150 - r_square: -0.0650 - val_loss: 6144.5376 - val_r_square: 0.2306\n",
      "Epoch 42/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11625.9229 - r_square: -0.0884 - val_loss: 8127.0947 - val_r_square: -0.0177\n",
      "Epoch 43/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 10994.3105 - r_square: -0.0293 - val_loss: 8063.1299 - val_r_square: -0.0097\n",
      "Epoch 44/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 11509.9600 - r_square: -0.0776 - val_loss: 8331.0967 - val_r_square: -0.0432\n",
      "Epoch 45/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11687.9756 - r_square: -0.0942 - val_loss: 8734.7705 - val_r_square: -0.0938\n",
      "Epoch 46/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11889.7510 - r_square: -0.1131 - val_loss: 9821.6709 - val_r_square: -0.2299\n",
      "Epoch 47/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11045.3018 - r_square: -0.0341 - val_loss: 7301.5586 - val_r_square: 0.0857\n",
      "Epoch 48/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11741.4229 - r_square: -0.0992 - val_loss: 7048.6021 - val_r_square: 0.1174\n",
      "Epoch 49/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11829.9932 - r_square: -0.1075 - val_loss: 6519.1294 - val_r_square: 0.1837\n",
      "Epoch 50/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 12112.1699 - r_square: -0.1340 - val_loss: 6492.6621 - val_r_square: 0.1870\n",
      "Epoch 51/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11421.7715 - r_square: -0.0693 - val_loss: 7253.4922 - val_r_square: 0.0917\n",
      "Epoch 52/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11864.1816 - r_square: -0.1107 - val_loss: 7248.9556 - val_r_square: 0.0923\n",
      "Epoch 53/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 10805.6621 - r_square: -0.0116 - val_loss: 8114.3398 - val_r_square: -0.0161\n",
      "Epoch 54/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11745.5869 - r_square: -0.0996 - val_loss: 8877.7568 - val_r_square: -0.1117\n",
      "Epoch 55/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11024.7842 - r_square: -0.0322 - val_loss: 7424.9365 - val_r_square: 0.0703\n",
      "Epoch 56/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 10401.1240 - r_square: 0.0262 - val_loss: 8431.6328 - val_r_square: -0.0558\n",
      "Epoch 57/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11899.6592 - r_square: -0.1141 - val_loss: 9585.7861 - val_r_square: -0.2003\n",
      "Epoch 58/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11660.4258 - r_square: -0.0917 - val_loss: 8189.7153 - val_r_square: -0.0255\n",
      "Epoch 59/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 10780.5742 - r_square: -0.0093 - val_loss: 7518.9912 - val_r_square: 0.0585\n",
      "Epoch 60/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11151.4189 - r_square: -0.0440 - val_loss: 6576.6182 - val_r_square: 0.1765\n",
      "Epoch 61/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 10622.4756 - r_square: 0.0055 - val_loss: 8610.8350 - val_r_square: -0.0782\n",
      "Epoch 62/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11030.3369 - r_square: -0.0327 - val_loss: 7430.7988 - val_r_square: 0.0695\n",
      "Epoch 63/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11354.6523 - r_square: -0.0630 - val_loss: 6570.6284 - val_r_square: 0.1772\n",
      "Epoch 64/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11392.3672 - r_square: -0.0666 - val_loss: 6137.5605 - val_r_square: 0.2315\n",
      "Epoch 65/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11444.5957 - r_square: -0.0715 - val_loss: 6598.1729 - val_r_square: 0.1738\n",
      "Epoch 66/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 12118.7344 - r_square: -0.1346 - val_loss: 11151.3262 - val_r_square: -0.3964\n",
      "Epoch 67/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 12860.4365 - r_square: -0.2040 - val_loss: 11933.7148 - val_r_square: -0.4943\n",
      "Epoch 68/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 13600.9043 - r_square: -0.2733 - val_loss: 11231.5469 - val_r_square: -0.4064\n",
      "Epoch 69/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 13481.1699 - r_square: -0.2621 - val_loss: 10480.9541 - val_r_square: -0.3124\n",
      "Epoch 70/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 12665.0557 - r_square: -0.1857 - val_loss: 8108.4653 - val_r_square: -0.0153\n",
      "Epoch 71/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 10928.9473 - r_square: -0.0232 - val_loss: 9447.7197 - val_r_square: -0.1830\n",
      "Epoch 72/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11745.2100 - r_square: -0.0996 - val_loss: 13384.6035 - val_r_square: -0.6760\n",
      "Epoch 73/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11715.9238 - r_square: -0.0969 - val_loss: 12406.9268 - val_r_square: -0.5536\n",
      "Epoch 74/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11521.3857 - r_square: -0.0787 - val_loss: 12050.2793 - val_r_square: -0.5089\n",
      "Epoch 75/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11271.6436 - r_square: -0.0553 - val_loss: 8633.6748 - val_r_square: -0.0811\n",
      "Epoch 76/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 10993.4189 - r_square: -0.0292 - val_loss: 6549.9272 - val_r_square: 0.1798\n",
      "Epoch 77/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 10763.1211 - r_square: -0.0077 - val_loss: 6412.9336 - val_r_square: 0.1970\n",
      "Epoch 78/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11582.6973 - r_square: -0.0844 - val_loss: 6700.4023 - val_r_square: 0.1610\n",
      "Epoch 79/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 12083.4707 - r_square: -0.1313 - val_loss: 8109.0781 - val_r_square: -0.0154\n",
      "Epoch 80/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11187.0840 - r_square: -0.0474 - val_loss: 7964.9912 - val_r_square: 0.0026\n",
      "Epoch 81/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11940.5742 - r_square: -0.1179 - val_loss: 7787.9150 - val_r_square: 0.0248\n",
      "Epoch 82/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11364.2354 - r_square: -0.0639 - val_loss: 13687.9238 - val_r_square: -0.7140\n",
      "Epoch 83/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11424.5586 - r_square: -0.0696 - val_loss: 9170.7461 - val_r_square: -0.1483\n",
      "Epoch 84/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11959.1768 - r_square: -0.1196 - val_loss: 11408.7295 - val_r_square: -0.4286\n",
      "Epoch 85/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11801.0273 - r_square: -0.1048 - val_loss: 13309.7188 - val_r_square: -0.6666\n",
      "Epoch 86/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 12242.1016 - r_square: -0.1461 - val_loss: 11132.6748 - val_r_square: -0.3940\n",
      "Epoch 87/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11336.1426 - r_square: -0.0613 - val_loss: 10899.5928 - val_r_square: -0.3648\n",
      "Epoch 88/400\n",
      "8/8 [==============================] - 0s 53ms/step - loss: 12216.1562 - r_square: -0.1437 - val_loss: 9111.8037 - val_r_square: -0.1410\n",
      "Epoch 89/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 11914.7051 - r_square: -0.1155 - val_loss: 6875.9600 - val_r_square: 0.1390\n",
      "Epoch 90/400\n",
      "8/8 [==============================] - 1s 77ms/step - loss: 11860.2832 - r_square: -0.1104 - val_loss: 7038.7900 - val_r_square: 0.1186\n",
      "Epoch 91/400\n",
      "8/8 [==============================] - 1s 66ms/step - loss: 11552.7773 - r_square: -0.0816 - val_loss: 7786.4780 - val_r_square: 0.0250\n",
      "Epoch 92/400\n",
      "8/8 [==============================] - 0s 60ms/step - loss: 11532.4365 - r_square: -0.0797 - val_loss: 7736.6025 - val_r_square: 0.0312\n",
      "Epoch 93/400\n",
      "8/8 [==============================] - 0s 61ms/step - loss: 11630.2725 - r_square: -0.0888 - val_loss: 6142.3457 - val_r_square: 0.2309\n",
      "Epoch 94/400\n",
      "8/8 [==============================] - 0s 53ms/step - loss: 11328.8262 - r_square: -0.0606 - val_loss: 5503.2852 - val_r_square: 0.3109\n",
      "Epoch 95/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 10850.0303 - r_square: -0.0158 - val_loss: 5661.7593 - val_r_square: 0.2910\n",
      "Epoch 96/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10747.8398 - r_square: -0.0062 - val_loss: 5612.4365 - val_r_square: 0.2972\n",
      "Epoch 97/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11479.2852 - r_square: -0.0747 - val_loss: 6046.0835 - val_r_square: 0.2429\n",
      "Epoch 98/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11099.5039 - r_square: -0.0391 - val_loss: 6146.8926 - val_r_square: 0.2303\n",
      "Epoch 99/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 10545.2803 - r_square: 0.0127 - val_loss: 6341.5547 - val_r_square: 0.2059\n",
      "Epoch 100/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 10727.1094 - r_square: -0.0043 - val_loss: 6473.3716 - val_r_square: 0.1894\n",
      "Epoch 101/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 10443.7832 - r_square: 0.0222 - val_loss: 6342.8208 - val_r_square: 0.2058\n",
      "Epoch 102/400\n",
      "8/8 [==============================] - 0s 51ms/step - loss: 11062.5674 - r_square: -0.0357 - val_loss: 6433.0493 - val_r_square: 0.1945\n",
      "Epoch 103/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 10989.9082 - r_square: -0.0289 - val_loss: 6803.9004 - val_r_square: 0.1480\n",
      "Epoch 104/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11320.4873 - r_square: -0.0598 - val_loss: 6795.8408 - val_r_square: 0.1490\n",
      "Epoch 105/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11366.5811 - r_square: -0.0642 - val_loss: 7230.1436 - val_r_square: 0.0947\n",
      "Epoch 106/400\n",
      "8/8 [==============================] - 0s 51ms/step - loss: 11179.5645 - r_square: -0.0466 - val_loss: 7729.1895 - val_r_square: 0.0322\n",
      "Epoch 107/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11082.0361 - r_square: -0.0375 - val_loss: 7142.5762 - val_r_square: 0.1056\n",
      "Epoch 108/400\n",
      "8/8 [==============================] - 0s 41ms/step - loss: 10825.0557 - r_square: -0.0135 - val_loss: 6774.9565 - val_r_square: 0.1517\n",
      "Epoch 109/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11315.5596 - r_square: -0.0594 - val_loss: 6165.7876 - val_r_square: 0.2279\n",
      "Epoch 110/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11548.4561 - r_square: -0.0812 - val_loss: 6376.5171 - val_r_square: 0.2015\n",
      "Epoch 111/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11603.6396 - r_square: -0.0863 - val_loss: 6602.1841 - val_r_square: 0.1733\n",
      "Epoch 112/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 10567.2979 - r_square: 0.0107 - val_loss: 6493.7603 - val_r_square: 0.1869\n",
      "Epoch 113/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11392.6436 - r_square: -0.0666 - val_loss: 6549.7827 - val_r_square: 0.1798\n",
      "Epoch 114/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11819.0947 - r_square: -0.1065 - val_loss: 6804.9897 - val_r_square: 0.1479\n",
      "Epoch 115/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11121.3154 - r_square: -0.0412 - val_loss: 5932.7969 - val_r_square: 0.2571\n",
      "Epoch 116/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11372.8242 - r_square: -0.0647 - val_loss: 5988.4370 - val_r_square: 0.2501\n",
      "Epoch 117/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10985.4873 - r_square: -0.0285 - val_loss: 6121.8667 - val_r_square: 0.2334\n",
      "Epoch 118/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11694.7354 - r_square: -0.0949 - val_loss: 5698.2656 - val_r_square: 0.2865\n",
      "Epoch 119/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10869.4092 - r_square: -0.0176 - val_loss: 5864.6270 - val_r_square: 0.2656\n",
      "Epoch 120/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10808.0488 - r_square: -0.0118 - val_loss: 5941.2437 - val_r_square: 0.2560\n",
      "Epoch 121/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11648.7100 - r_square: -0.0906 - val_loss: 6291.3086 - val_r_square: 0.2122\n",
      "Epoch 122/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11203.7041 - r_square: -0.0489 - val_loss: 6389.5796 - val_r_square: 0.1999\n",
      "Epoch 123/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11222.4492 - r_square: -0.0507 - val_loss: 6495.2832 - val_r_square: 0.1867\n",
      "Epoch 124/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11708.6064 - r_square: -0.0962 - val_loss: 6473.8491 - val_r_square: 0.1894\n",
      "Epoch 125/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10850.6602 - r_square: -0.0159 - val_loss: 6316.3433 - val_r_square: 0.2091\n",
      "Epoch 126/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 10556.9873 - r_square: 0.0116 - val_loss: 6117.9282 - val_r_square: 0.2339\n",
      "Epoch 127/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11682.6631 - r_square: -0.0937 - val_loss: 5837.3813 - val_r_square: 0.2691\n",
      "Epoch 128/400\n",
      "8/8 [==============================] - 0s 41ms/step - loss: 11599.9141 - r_square: -0.0860 - val_loss: 6166.7739 - val_r_square: 0.2278\n",
      "Epoch 129/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 10724.0312 - r_square: -0.0040 - val_loss: 5747.3599 - val_r_square: 0.2803\n",
      "Epoch 130/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11043.3760 - r_square: -0.0339 - val_loss: 5881.0542 - val_r_square: 0.2636\n",
      "Epoch 131/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11074.5352 - r_square: -0.0368 - val_loss: 5865.9214 - val_r_square: 0.2655\n",
      "Epoch 132/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11234.2393 - r_square: -0.0518 - val_loss: 6098.8979 - val_r_square: 0.2363\n",
      "Epoch 133/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12118.8223 - r_square: -0.1346 - val_loss: 6299.6167 - val_r_square: 0.2112\n",
      "Epoch 134/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11170.7412 - r_square: -0.0458 - val_loss: 6214.8999 - val_r_square: 0.2218\n",
      "Epoch 135/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11984.5977 - r_square: -0.1220 - val_loss: 6398.8608 - val_r_square: 0.1987\n",
      "Epoch 136/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 10935.1318 - r_square: -0.0238 - val_loss: 6572.7588 - val_r_square: 0.1770\n",
      "Epoch 137/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11119.1885 - r_square: -0.0410 - val_loss: 6146.3730 - val_r_square: 0.2304\n",
      "Epoch 138/400\n",
      "8/8 [==============================] - 0s 51ms/step - loss: 11369.1768 - r_square: -0.0644 - val_loss: 5906.4478 - val_r_square: 0.2604\n",
      "Epoch 139/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 10930.3125 - r_square: -0.0233 - val_loss: 5944.7690 - val_r_square: 0.2556\n",
      "Epoch 140/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11083.6152 - r_square: -0.0377 - val_loss: 5854.4707 - val_r_square: 0.2669\n",
      "Epoch 141/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11332.1699 - r_square: -0.0609 - val_loss: 5995.1587 - val_r_square: 0.2493\n",
      "Epoch 142/400\n",
      "8/8 [==============================] - 0s 51ms/step - loss: 11449.6309 - r_square: -0.0719 - val_loss: 5865.9033 - val_r_square: 0.2655\n",
      "Epoch 143/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11280.3428 - r_square: -0.0561 - val_loss: 5947.3618 - val_r_square: 0.2553\n",
      "Epoch 144/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11369.4512 - r_square: -0.0644 - val_loss: 5639.4141 - val_r_square: 0.2938\n",
      "Epoch 145/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11751.5439 - r_square: -0.1002 - val_loss: 5500.4058 - val_r_square: 0.3112\n",
      "Epoch 146/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11225.6641 - r_square: -0.0510 - val_loss: 5679.3315 - val_r_square: 0.2888\n",
      "Epoch 147/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10618.3770 - r_square: 0.0059 - val_loss: 5473.0547 - val_r_square: 0.3147\n",
      "Epoch 148/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11652.9414 - r_square: -0.0910 - val_loss: 5599.7852 - val_r_square: 0.2988\n",
      "Epoch 149/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11397.7646 - r_square: -0.0671 - val_loss: 5783.4272 - val_r_square: 0.2758\n",
      "Epoch 150/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11198.6543 - r_square: -0.0484 - val_loss: 5816.5957 - val_r_square: 0.2717\n",
      "Epoch 151/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 10803.3525 - r_square: -0.0114 - val_loss: 5581.6934 - val_r_square: 0.3011\n",
      "Epoch 152/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 10695.1621 - r_square: -0.0013 - val_loss: 5475.0703 - val_r_square: 0.3144\n",
      "Epoch 153/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 10634.8262 - r_square: 0.0044 - val_loss: 5862.2485 - val_r_square: 0.2659\n",
      "Epoch 154/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11127.7080 - r_square: -0.0418 - val_loss: 5793.0596 - val_r_square: 0.2746\n",
      "Epoch 155/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11915.4043 - r_square: -0.1155 - val_loss: 5890.2607 - val_r_square: 0.2624\n",
      "Epoch 156/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11391.9805 - r_square: -0.0665 - val_loss: 6186.0293 - val_r_square: 0.2254\n",
      "Epoch 157/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11488.4922 - r_square: -0.0756 - val_loss: 6449.3672 - val_r_square: 0.1924\n",
      "Epoch 158/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11273.8828 - r_square: -0.0555 - val_loss: 6484.4487 - val_r_square: 0.1880\n",
      "Epoch 159/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11283.1689 - r_square: -0.0563 - val_loss: 6857.4922 - val_r_square: 0.1413\n",
      "Epoch 160/400\n",
      "8/8 [==============================] - 0s 58ms/step - loss: 10842.1582 - r_square: -0.0151 - val_loss: 6518.0190 - val_r_square: 0.1838\n",
      "Epoch 161/400\n",
      "8/8 [==============================] - 0s 60ms/step - loss: 10753.4971 - r_square: -0.0068 - val_loss: 6049.4463 - val_r_square: 0.2425\n",
      "Epoch 162/400\n",
      "8/8 [==============================] - 0s 62ms/step - loss: 11511.1064 - r_square: -0.0777 - val_loss: 6090.0190 - val_r_square: 0.2374\n",
      "Epoch 163/400\n",
      "8/8 [==============================] - 1s 72ms/step - loss: 11298.8838 - r_square: -0.0578 - val_loss: 6211.4165 - val_r_square: 0.2222\n",
      "Epoch 164/400\n",
      "8/8 [==============================] - 0s 54ms/step - loss: 10893.3174 - r_square: -0.0199 - val_loss: 6315.4692 - val_r_square: 0.2092\n",
      "Epoch 165/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11478.9111 - r_square: -0.0747 - val_loss: 6332.0762 - val_r_square: 0.2071\n",
      "Epoch 166/400\n",
      "8/8 [==============================] - 0s 57ms/step - loss: 10576.7539 - r_square: 0.0098 - val_loss: 6177.4604 - val_r_square: 0.2265\n",
      "Epoch 167/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11418.8779 - r_square: -0.0690 - val_loss: 6145.8940 - val_r_square: 0.2304\n",
      "Epoch 168/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11702.8457 - r_square: -0.0956 - val_loss: 5692.4814 - val_r_square: 0.2872\n",
      "Epoch 169/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11540.2910 - r_square: -0.0804 - val_loss: 5813.4263 - val_r_square: 0.2721\n",
      "Epoch 170/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 12057.3643 - r_square: -0.1288 - val_loss: 5477.4771 - val_r_square: 0.3141\n",
      "Epoch 171/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11881.7070 - r_square: -0.1124 - val_loss: 6537.1943 - val_r_square: 0.1814\n",
      "Epoch 172/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11192.3320 - r_square: -0.0478 - val_loss: 7035.1870 - val_r_square: 0.1191\n",
      "Epoch 173/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11042.3838 - r_square: -0.0338 - val_loss: 6892.5942 - val_r_square: 0.1369\n",
      "Epoch 174/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11416.8672 - r_square: -0.0689 - val_loss: 6055.2407 - val_r_square: 0.2418\n",
      "Epoch 175/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11278.9570 - r_square: -0.0559 - val_loss: 6724.8208 - val_r_square: 0.1579\n",
      "Epoch 176/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11704.9160 - r_square: -0.0958 - val_loss: 6707.5420 - val_r_square: 0.1601\n",
      "Epoch 177/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 10688.4043 - r_square: -6.5458e-04 - val_loss: 7365.6841 - val_r_square: 0.0777\n",
      "Epoch 178/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11030.2002 - r_square: -0.0327 - val_loss: 6722.3545 - val_r_square: 0.1582\n",
      "Epoch 179/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11723.7852 - r_square: -0.0976 - val_loss: 6749.5430 - val_r_square: 0.1548\n",
      "Epoch 180/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11303.0938 - r_square: -0.0582 - val_loss: 6426.4277 - val_r_square: 0.1953\n",
      "Epoch 181/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 10340.2979 - r_square: 0.0319 - val_loss: 6552.1890 - val_r_square: 0.1795\n",
      "Epoch 182/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 10371.3281 - r_square: 0.0290 - val_loss: 7779.9492 - val_r_square: 0.0258\n",
      "Epoch 183/400\n",
      "8/8 [==============================] - 1s 63ms/step - loss: 11804.2627 - r_square: -0.1051 - val_loss: 9368.0801 - val_r_square: -0.1731\n",
      "Epoch 184/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 10694.0752 - r_square: -0.0012 - val_loss: 6598.7598 - val_r_square: 0.1737\n",
      "Epoch 185/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10487.3652 - r_square: 0.0182 - val_loss: 6100.2803 - val_r_square: 0.2361\n",
      "Epoch 186/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11621.2441 - r_square: -0.0880 - val_loss: 6281.2075 - val_r_square: 0.2135\n",
      "Epoch 187/400\n",
      "8/8 [==============================] - 0s 58ms/step - loss: 11607.5352 - r_square: -0.0867 - val_loss: 6153.8354 - val_r_square: 0.2294\n",
      "Epoch 188/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 10881.7148 - r_square: -0.0188 - val_loss: 6252.8555 - val_r_square: 0.2170\n",
      "Epoch 189/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 11873.0176 - r_square: -0.1116 - val_loss: 6153.8853 - val_r_square: 0.2294\n",
      "Epoch 190/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 10872.5830 - r_square: -0.0179 - val_loss: 6357.0249 - val_r_square: 0.2040\n",
      "Epoch 191/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11200.6455 - r_square: -0.0486 - val_loss: 6174.2275 - val_r_square: 0.2269\n",
      "Epoch 192/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11207.7344 - r_square: -0.0493 - val_loss: 6013.3940 - val_r_square: 0.2470\n",
      "Epoch 193/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 10856.9277 - r_square: -0.0164 - val_loss: 5850.7773 - val_r_square: 0.2674\n",
      "Epoch 194/400\n",
      "8/8 [==============================] - 0s 65ms/step - loss: 10294.6611 - r_square: 0.0362 - val_loss: 6518.4121 - val_r_square: 0.1838\n",
      "Epoch 195/400\n",
      "8/8 [==============================] - 0s 41ms/step - loss: 11315.5107 - r_square: -0.0594 - val_loss: 6870.5342 - val_r_square: 0.1397\n",
      "Epoch 196/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10282.4404 - r_square: 0.0374 - val_loss: 6750.6367 - val_r_square: 0.1547\n",
      "Epoch 197/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10711.5449 - r_square: -0.0028 - val_loss: 6248.5435 - val_r_square: 0.2176\n",
      "Epoch 198/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10205.0596 - r_square: 0.0446 - val_loss: 6736.4536 - val_r_square: 0.1565\n",
      "Epoch 199/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 10913.5342 - r_square: -0.0217 - val_loss: 6964.4023 - val_r_square: 0.1279\n",
      "Epoch 200/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10731.9883 - r_square: -0.0047 - val_loss: 5977.1616 - val_r_square: 0.2515\n",
      "Epoch 201/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10629.9727 - r_square: 0.0048 - val_loss: 5704.5771 - val_r_square: 0.2857\n",
      "Epoch 202/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11262.3477 - r_square: -0.0544 - val_loss: 5736.7227 - val_r_square: 0.2817\n",
      "Epoch 203/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11231.4854 - r_square: -0.0515 - val_loss: 6091.6958 - val_r_square: 0.2372\n",
      "Epoch 204/400\n",
      "8/8 [==============================] - 1s 66ms/step - loss: 11396.1504 - r_square: -0.0669 - val_loss: 5941.5967 - val_r_square: 0.2560\n",
      "Epoch 205/400\n",
      "8/8 [==============================] - 0s 65ms/step - loss: 10878.6172 - r_square: -0.0185 - val_loss: 5549.9092 - val_r_square: 0.3050\n",
      "Epoch 206/400\n",
      "8/8 [==============================] - 0s 64ms/step - loss: 11203.1240 - r_square: -0.0488 - val_loss: 5691.5913 - val_r_square: 0.2873\n",
      "Epoch 207/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 10878.5791 - r_square: -0.0185 - val_loss: 6549.1919 - val_r_square: 0.1799\n",
      "Epoch 208/400\n",
      "8/8 [==============================] - 0s 53ms/step - loss: 10601.9746 - r_square: 0.0074 - val_loss: 6017.4473 - val_r_square: 0.2465\n",
      "Epoch 209/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 12069.6494 - r_square: -0.1300 - val_loss: 5785.1655 - val_r_square: 0.2756\n",
      "Epoch 210/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10293.1279 - r_square: 0.0363 - val_loss: 5916.4033 - val_r_square: 0.2592\n",
      "Epoch 211/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 10919.9492 - r_square: -0.0223 - val_loss: 5688.6470 - val_r_square: 0.2877\n",
      "Epoch 212/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 9560.3994 - r_square: 0.1049 - val_loss: 5873.2915 - val_r_square: 0.2646\n",
      "Epoch 213/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11296.6445 - r_square: -0.0576 - val_loss: 6143.6001 - val_r_square: 0.2307\n",
      "Epoch 214/400\n",
      "8/8 [==============================] - 0s 62ms/step - loss: 10535.0957 - r_square: 0.0137 - val_loss: 5774.1392 - val_r_square: 0.2770\n",
      "Epoch 215/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11243.7764 - r_square: -0.0527 - val_loss: 5564.9712 - val_r_square: 0.3032\n",
      "Epoch 216/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 10996.3857 - r_square: -0.0295 - val_loss: 5666.0957 - val_r_square: 0.2905\n",
      "Epoch 217/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11580.8096 - r_square: -0.0842 - val_loss: 5603.0171 - val_r_square: 0.2984\n",
      "Epoch 218/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11111.3115 - r_square: -0.0403 - val_loss: 5535.8267 - val_r_square: 0.3068\n",
      "Epoch 219/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11032.4697 - r_square: -0.0329 - val_loss: 5360.9941 - val_r_square: 0.3287\n",
      "Epoch 220/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11319.9512 - r_square: -0.0598 - val_loss: 5706.7271 - val_r_square: 0.2854\n",
      "Epoch 221/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 11288.0537 - r_square: -0.0568 - val_loss: 7026.8140 - val_r_square: 0.1201\n",
      "Epoch 222/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 10723.0918 - r_square: -0.0039 - val_loss: 8478.8662 - val_r_square: -0.0617\n",
      "Epoch 223/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11570.3408 - r_square: -0.0832 - val_loss: 6959.4414 - val_r_square: 0.1286\n",
      "Epoch 224/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11068.8027 - r_square: -0.0363 - val_loss: 7419.0513 - val_r_square: 0.0710\n",
      "Epoch 225/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11549.9502 - r_square: -0.0813 - val_loss: 7619.4644 - val_r_square: 0.0459\n",
      "Epoch 226/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11312.8750 - r_square: -0.0591 - val_loss: 6932.3745 - val_r_square: 0.1319\n",
      "Epoch 227/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 10673.5322 - r_square: 7.3111e-04 - val_loss: 6505.9443 - val_r_square: 0.1853\n",
      "Epoch 228/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11158.4619 - r_square: -0.0447 - val_loss: 6366.4946 - val_r_square: 0.2028\n",
      "Epoch 229/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11056.4375 - r_square: -0.0351 - val_loss: 7402.6943 - val_r_square: 0.0730\n",
      "Epoch 230/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10949.3086 - r_square: -0.0251 - val_loss: 6100.4941 - val_r_square: 0.2361\n",
      "Epoch 231/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10730.0400 - r_square: -0.0046 - val_loss: 6239.8955 - val_r_square: 0.2187\n",
      "Epoch 232/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11244.8643 - r_square: -0.0528 - val_loss: 5791.1348 - val_r_square: 0.2748\n",
      "Epoch 233/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11008.4893 - r_square: -0.0306 - val_loss: 6290.4028 - val_r_square: 0.2123\n",
      "Epoch 234/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 10980.8604 - r_square: -0.0280 - val_loss: 6447.0859 - val_r_square: 0.1927\n",
      "Epoch 235/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11935.0107 - r_square: -0.1174 - val_loss: 7466.8218 - val_r_square: 0.0650\n",
      "Epoch 236/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11111.4346 - r_square: -0.0403 - val_loss: 19465.4355 - val_r_square: -1.4374\n",
      "Epoch 237/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 10822.7764 - r_square: -0.0132 - val_loss: 20455.7637 - val_r_square: -1.5614\n",
      "Epoch 238/400\n",
      "8/8 [==============================] - 0s 61ms/step - loss: 11695.1211 - r_square: -0.0949 - val_loss: 11274.6904 - val_r_square: -0.4118\n",
      "Epoch 239/400\n",
      "8/8 [==============================] - 1s 66ms/step - loss: 12923.8545 - r_square: -0.2100 - val_loss: 8438.2793 - val_r_square: -0.0566\n",
      "Epoch 240/400\n",
      "8/8 [==============================] - 0s 60ms/step - loss: 12269.6826 - r_square: -0.1487 - val_loss: 7521.4858 - val_r_square: 0.0582\n",
      "Epoch 241/400\n",
      "8/8 [==============================] - 0s 56ms/step - loss: 12194.1328 - r_square: -0.1416 - val_loss: 7859.2627 - val_r_square: 0.0159\n",
      "Epoch 242/400\n",
      "8/8 [==============================] - 0s 55ms/step - loss: 12165.8643 - r_square: -0.1390 - val_loss: 8096.9067 - val_r_square: -0.0139\n",
      "Epoch 243/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11237.6260 - r_square: -0.0521 - val_loss: 7684.4673 - val_r_square: 0.0378\n",
      "Epoch 244/400\n",
      "8/8 [==============================] - 0s 56ms/step - loss: 11817.8223 - r_square: -0.1064 - val_loss: 7608.1284 - val_r_square: 0.0473\n",
      "Epoch 245/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12312.1670 - r_square: -0.1527 - val_loss: 6386.5366 - val_r_square: 0.2003\n",
      "Epoch 246/400\n",
      "8/8 [==============================] - 0s 51ms/step - loss: 11956.3809 - r_square: -0.1194 - val_loss: 6971.6138 - val_r_square: 0.1270\n",
      "Epoch 247/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 12991.5771 - r_square: -0.2163 - val_loss: 7871.9922 - val_r_square: 0.0143\n",
      "Epoch 248/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12038.3848 - r_square: -0.1271 - val_loss: 8505.9404 - val_r_square: -0.0651\n",
      "Epoch 249/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12662.5508 - r_square: -0.1855 - val_loss: 8435.8545 - val_r_square: -0.0563\n",
      "Epoch 250/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12029.4424 - r_square: -0.1262 - val_loss: 7479.8794 - val_r_square: 0.0634\n",
      "Epoch 251/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11398.2559 - r_square: -0.0671 - val_loss: 7940.2686 - val_r_square: 0.0057\n",
      "Epoch 252/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12074.6289 - r_square: -0.1304 - val_loss: 6911.8945 - val_r_square: 0.1345\n",
      "Epoch 253/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11826.6689 - r_square: -0.1072 - val_loss: 7611.3130 - val_r_square: 0.0469\n",
      "Epoch 254/400\n",
      "8/8 [==============================] - 0s 53ms/step - loss: 11495.1328 - r_square: -0.0762 - val_loss: 7386.9980 - val_r_square: 0.0750\n",
      "Epoch 255/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 11417.8340 - r_square: -0.0689 - val_loss: 6259.4165 - val_r_square: 0.2162\n",
      "Epoch 256/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11761.2461 - r_square: -0.1011 - val_loss: 5890.1704 - val_r_square: 0.2624\n",
      "Epoch 257/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11634.4736 - r_square: -0.0892 - val_loss: 6152.8623 - val_r_square: 0.2295\n",
      "Epoch 258/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11972.6299 - r_square: -0.1209 - val_loss: 6059.6641 - val_r_square: 0.2412\n",
      "Epoch 259/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12016.1455 - r_square: -0.1250 - val_loss: 6004.4512 - val_r_square: 0.2481\n",
      "Epoch 260/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10843.1875 - r_square: -0.0152 - val_loss: 6093.0435 - val_r_square: 0.2370\n",
      "Epoch 261/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11428.3232 - r_square: -0.0699 - val_loss: 5846.4316 - val_r_square: 0.2679\n",
      "Epoch 262/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11276.3613 - r_square: -0.0557 - val_loss: 6147.4023 - val_r_square: 0.2302\n",
      "Epoch 263/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11437.3770 - r_square: -0.0708 - val_loss: 5875.1450 - val_r_square: 0.2643\n",
      "Epoch 264/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12129.6299 - r_square: -0.1356 - val_loss: 6065.1758 - val_r_square: 0.2405\n",
      "Epoch 265/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12027.0029 - r_square: -0.1260 - val_loss: 5913.6763 - val_r_square: 0.2595\n",
      "Epoch 266/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11491.7646 - r_square: -0.0759 - val_loss: 6177.0293 - val_r_square: 0.2265\n",
      "Epoch 267/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11559.3320 - r_square: -0.0822 - val_loss: 6527.8184 - val_r_square: 0.1826\n",
      "Epoch 268/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11399.4023 - r_square: -0.0672 - val_loss: 7637.8613 - val_r_square: 0.0436\n",
      "Epoch 269/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 10566.0000 - r_square: 0.0108 - val_loss: 7381.0981 - val_r_square: 0.0758\n",
      "Epoch 270/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11431.6504 - r_square: -0.0702 - val_loss: 7266.6133 - val_r_square: 0.0901\n",
      "Epoch 271/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11669.1348 - r_square: -0.0925 - val_loss: 6805.9673 - val_r_square: 0.1478\n",
      "Epoch 272/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11194.2695 - r_square: -0.0480 - val_loss: 6913.2363 - val_r_square: 0.1343\n",
      "Epoch 273/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12063.4023 - r_square: -0.1294 - val_loss: 8710.7334 - val_r_square: -0.0907\n",
      "Epoch 274/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 12380.1787 - r_square: -0.1591 - val_loss: 8231.7520 - val_r_square: -0.0308\n",
      "Epoch 275/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11202.5322 - r_square: -0.0488 - val_loss: 7805.3960 - val_r_square: 0.0226\n",
      "Epoch 276/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11682.4824 - r_square: -0.0937 - val_loss: 6225.6357 - val_r_square: 0.2204\n",
      "Epoch 277/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12054.3203 - r_square: -0.1285 - val_loss: 7072.6870 - val_r_square: 0.1144\n",
      "Epoch 278/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12521.4902 - r_square: -0.1723 - val_loss: 8520.0166 - val_r_square: -0.0669\n",
      "Epoch 279/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12777.6963 - r_square: -0.1963 - val_loss: 7994.7998 - val_r_square: -0.0011\n",
      "Epoch 280/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 13203.5820 - r_square: -0.2361 - val_loss: 7545.7085 - val_r_square: 0.0551\n",
      "Epoch 281/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12535.4307 - r_square: -0.1736 - val_loss: 7327.7510 - val_r_square: 0.0824\n",
      "Epoch 282/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12517.7051 - r_square: -0.1719 - val_loss: 7241.4424 - val_r_square: 0.0932\n",
      "Epoch 283/400\n",
      "8/8 [==============================] - 0s 53ms/step - loss: 11713.1807 - r_square: -0.0966 - val_loss: 7626.0864 - val_r_square: 0.0451\n",
      "Epoch 284/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 12437.9941 - r_square: -0.1645 - val_loss: 7386.6924 - val_r_square: 0.0751\n",
      "Epoch 285/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 12354.1221 - r_square: -0.1566 - val_loss: 7323.6958 - val_r_square: 0.0829\n",
      "Epoch 286/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 12421.1191 - r_square: -0.1629 - val_loss: 6509.5186 - val_r_square: 0.1849\n",
      "Epoch 287/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11002.1045 - r_square: -0.0300 - val_loss: 6140.3608 - val_r_square: 0.2311\n",
      "Epoch 288/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11993.5664 - r_square: -0.1229 - val_loss: 5858.3989 - val_r_square: 0.2664\n",
      "Epoch 289/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11766.2051 - r_square: -0.1016 - val_loss: 5877.9175 - val_r_square: 0.2640\n",
      "Epoch 290/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11864.5439 - r_square: -0.1108 - val_loss: 5922.0498 - val_r_square: 0.2585\n",
      "Epoch 291/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11470.6221 - r_square: -0.0739 - val_loss: 6028.4326 - val_r_square: 0.2451\n",
      "Epoch 292/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11034.1973 - r_square: -0.0330 - val_loss: 6308.6904 - val_r_square: 0.2100\n",
      "Epoch 293/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11176.2197 - r_square: -0.0463 - val_loss: 6291.4756 - val_r_square: 0.2122\n",
      "Epoch 294/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11276.4502 - r_square: -0.0557 - val_loss: 6425.0249 - val_r_square: 0.1955\n",
      "Epoch 295/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11467.4551 - r_square: -0.0736 - val_loss: 6468.3384 - val_r_square: 0.1900\n",
      "Epoch 296/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11240.0488 - r_square: -0.0523 - val_loss: 6494.4214 - val_r_square: 0.1868\n",
      "Epoch 297/400\n",
      "8/8 [==============================] - 0s 53ms/step - loss: 11738.1162 - r_square: -0.0989 - val_loss: 6533.2695 - val_r_square: 0.1819\n",
      "Epoch 298/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11373.7070 - r_square: -0.0648 - val_loss: 6803.9790 - val_r_square: 0.1480\n",
      "Epoch 299/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 11353.7744 - r_square: -0.0630 - val_loss: 6714.9248 - val_r_square: 0.1592\n",
      "Epoch 300/400\n",
      "8/8 [==============================] - 0s 61ms/step - loss: 11098.3223 - r_square: -0.0390 - val_loss: 6420.5127 - val_r_square: 0.1960\n",
      "Epoch 301/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 11042.1641 - r_square: -0.0338 - val_loss: 6097.7612 - val_r_square: 0.2364\n",
      "Epoch 302/400\n",
      "8/8 [==============================] - 0s 53ms/step - loss: 11437.1562 - r_square: -0.0708 - val_loss: 5990.2451 - val_r_square: 0.2499\n",
      "Epoch 303/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 11884.7334 - r_square: -0.1127 - val_loss: 6678.0908 - val_r_square: 0.1638\n",
      "Epoch 304/400\n",
      "8/8 [==============================] - 0s 51ms/step - loss: 11571.5459 - r_square: -0.0833 - val_loss: 6793.6631 - val_r_square: 0.1493\n",
      "Epoch 305/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 12388.7393 - r_square: -0.1599 - val_loss: 5940.9150 - val_r_square: 0.2561\n",
      "Epoch 306/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11721.6621 - r_square: -0.0974 - val_loss: 6633.7500 - val_r_square: 0.1693\n",
      "Epoch 307/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11739.0283 - r_square: -0.0990 - val_loss: 6753.4458 - val_r_square: 0.1543\n",
      "Epoch 308/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11741.8555 - r_square: -0.0993 - val_loss: 6509.0498 - val_r_square: 0.1849\n",
      "Epoch 309/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 12604.7441 - r_square: -0.1801 - val_loss: 14947.2432 - val_r_square: -0.8717\n",
      "Epoch 310/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 11603.9453 - r_square: -0.0864 - val_loss: 13154.0127 - val_r_square: -0.6471\n",
      "Epoch 311/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 12508.5225 - r_square: -0.1711 - val_loss: 9580.0986 - val_r_square: -0.1996\n",
      "Epoch 312/400\n",
      "8/8 [==============================] - 0s 54ms/step - loss: 12656.5889 - r_square: -0.1849 - val_loss: 11946.6758 - val_r_square: -0.4959\n",
      "Epoch 313/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 12631.8584 - r_square: -0.1826 - val_loss: 17112.5215 - val_r_square: -1.1428\n",
      "Epoch 314/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 13389.5957 - r_square: -0.2535 - val_loss: 10021.8281 - val_r_square: -0.2549\n",
      "Epoch 315/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 12041.6992 - r_square: -0.1274 - val_loss: 9536.8662 - val_r_square: -0.1942\n",
      "Epoch 316/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 11837.0547 - r_square: -0.1082 - val_loss: 10211.7207 - val_r_square: -0.2787\n",
      "Epoch 317/400\n",
      "8/8 [==============================] - 0s 57ms/step - loss: 12704.8955 - r_square: -0.1895 - val_loss: 9052.9287 - val_r_square: -0.1336\n",
      "Epoch 318/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 13368.4443 - r_square: -0.2516 - val_loss: 8056.9727 - val_r_square: -0.0089\n",
      "Epoch 319/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 12388.1875 - r_square: -0.1598 - val_loss: 9037.2617 - val_r_square: -0.1316\n",
      "Epoch 320/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 12172.7480 - r_square: -0.1396 - val_loss: 9124.5479 - val_r_square: -0.1426\n",
      "Epoch 321/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11863.5869 - r_square: -0.1107 - val_loss: 12379.8633 - val_r_square: -0.5502\n",
      "Epoch 322/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11800.8281 - r_square: -0.1048 - val_loss: 7667.1392 - val_r_square: 0.0399\n",
      "Epoch 323/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12118.1777 - r_square: -0.1345 - val_loss: 7811.1118 - val_r_square: 0.0219\n",
      "Epoch 324/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11814.0508 - r_square: -0.1060 - val_loss: 7367.5947 - val_r_square: 0.0774\n",
      "Epoch 325/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11759.6348 - r_square: -0.1010 - val_loss: 8952.6982 - val_r_square: -0.1210\n",
      "Epoch 326/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11464.4248 - r_square: -0.0733 - val_loss: 7097.5576 - val_r_square: 0.1113\n",
      "Epoch 327/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 11792.4160 - r_square: -0.1040 - val_loss: 7129.0903 - val_r_square: 0.1073\n",
      "Epoch 328/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 11721.1250 - r_square: -0.0973 - val_loss: 6379.2534 - val_r_square: 0.2012\n",
      "Epoch 329/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11455.5674 - r_square: -0.0725 - val_loss: 8848.3506 - val_r_square: -0.1080\n",
      "Epoch 330/400\n",
      "8/8 [==============================] - 0s 41ms/step - loss: 12398.2109 - r_square: -0.1607 - val_loss: 8601.8857 - val_r_square: -0.0771\n",
      "Epoch 331/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 12784.8223 - r_square: -0.1969 - val_loss: 8248.9492 - val_r_square: -0.0329\n",
      "Epoch 332/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 13151.7959 - r_square: -0.2313 - val_loss: 72872.0312 - val_r_square: -8.1249\n",
      "Epoch 333/400\n",
      "8/8 [==============================] - 0s 44ms/step - loss: 12414.0449 - r_square: -0.1622 - val_loss: 101482.7109 - val_r_square: -11.7075\n",
      "Epoch 334/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 12238.8193 - r_square: -0.1458 - val_loss: 100261.6719 - val_r_square: -11.5546\n",
      "Epoch 335/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 13313.1719 - r_square: -0.2464 - val_loss: 78195.5547 - val_r_square: -8.7915\n",
      "Epoch 336/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11993.4834 - r_square: -0.1228 - val_loss: 83037.6016 - val_r_square: -9.3978\n",
      "Epoch 337/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12435.7197 - r_square: -0.1643 - val_loss: 79035.3984 - val_r_square: -8.8967\n",
      "Epoch 338/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12277.3809 - r_square: -0.1494 - val_loss: 75270.2969 - val_r_square: -8.4252\n",
      "Epoch 339/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11427.0039 - r_square: -0.0698 - val_loss: 73903.7031 - val_r_square: -8.2541\n",
      "Epoch 340/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12646.7383 - r_square: -0.1840 - val_loss: 69164.8906 - val_r_square: -7.6607\n",
      "Epoch 341/400\n",
      "8/8 [==============================] - 0s 48ms/step - loss: 12347.9355 - r_square: -0.1560 - val_loss: 55645.3750 - val_r_square: -5.9678\n",
      "Epoch 342/400\n",
      "8/8 [==============================] - 0s 51ms/step - loss: 12881.9941 - r_square: -0.2060 - val_loss: 48737.8789 - val_r_square: -5.1029\n",
      "Epoch 343/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 11919.0781 - r_square: -0.1159 - val_loss: 35042.9141 - val_r_square: -3.3880\n",
      "Epoch 344/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12399.7568 - r_square: -0.1609 - val_loss: 30680.7246 - val_r_square: -2.8418\n",
      "Epoch 345/400\n",
      "8/8 [==============================] - 0s 52ms/step - loss: 13064.5020 - r_square: -0.2231 - val_loss: 16262.0322 - val_r_square: -1.0363\n",
      "Epoch 346/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12639.7285 - r_square: -0.1834 - val_loss: 11913.6709 - val_r_square: -0.4918\n",
      "Epoch 347/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 12027.8506 - r_square: -0.1261 - val_loss: 11175.8555 - val_r_square: -0.3994\n",
      "Epoch 348/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 12530.2227 - r_square: -0.1731 - val_loss: 11528.8281 - val_r_square: -0.4436\n",
      "Epoch 349/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12263.9248 - r_square: -0.1482 - val_loss: 11287.1465 - val_r_square: -0.4134\n",
      "Epoch 350/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12395.9785 - r_square: -0.1605 - val_loss: 11138.1191 - val_r_square: -0.3947\n",
      "Epoch 351/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12577.6523 - r_square: -0.1775 - val_loss: 11229.8818 - val_r_square: -0.4062\n",
      "Epoch 352/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11770.6709 - r_square: -0.1020 - val_loss: 11429.9043 - val_r_square: -0.4312\n",
      "Epoch 353/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12387.5342 - r_square: -0.1597 - val_loss: 18325.3457 - val_r_square: -1.2947\n",
      "Epoch 354/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12681.6025 - r_square: -0.1873 - val_loss: 9817.6172 - val_r_square: -0.2293\n",
      "Epoch 355/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12806.7451 - r_square: -0.1990 - val_loss: 13292.1504 - val_r_square: -0.6644\n",
      "Epoch 356/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12073.2539 - r_square: -0.1303 - val_loss: 10888.9648 - val_r_square: -0.3635\n",
      "Epoch 357/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10789.1904 - r_square: -0.0101 - val_loss: 12759.1904 - val_r_square: -0.5977\n",
      "Epoch 358/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12336.4463 - r_square: -0.1549 - val_loss: 12116.0645 - val_r_square: -0.5172\n",
      "Epoch 359/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12689.2734 - r_square: -0.1880 - val_loss: 11497.7764 - val_r_square: -0.4397\n",
      "Epoch 360/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11917.2070 - r_square: -0.1157 - val_loss: 11461.3047 - val_r_square: -0.4352\n",
      "Epoch 361/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11649.7656 - r_square: -0.0907 - val_loss: 6767.3647 - val_r_square: 0.1526\n",
      "Epoch 362/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12192.8760 - r_square: -0.1415 - val_loss: 6945.8472 - val_r_square: 0.1303\n",
      "Epoch 363/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11419.8018 - r_square: -0.0691 - val_loss: 7529.8301 - val_r_square: 0.0571\n",
      "Epoch 364/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12397.8438 - r_square: -0.1607 - val_loss: 10039.5977 - val_r_square: -0.2571\n",
      "Epoch 365/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 12660.3340 - r_square: -0.1853 - val_loss: 11765.5684 - val_r_square: -0.4733\n",
      "Epoch 366/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10935.8330 - r_square: -0.0238 - val_loss: 8742.6250 - val_r_square: -0.0947\n",
      "Epoch 367/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12092.3213 - r_square: -0.1321 - val_loss: 7771.2607 - val_r_square: 0.0269\n",
      "Epoch 368/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11948.7842 - r_square: -0.1187 - val_loss: 7276.3643 - val_r_square: 0.0889\n",
      "Epoch 369/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 12184.6514 - r_square: -0.1407 - val_loss: 7436.4756 - val_r_square: 0.0688\n",
      "Epoch 370/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12292.5684 - r_square: -0.1508 - val_loss: 8658.3311 - val_r_square: -0.0842\n",
      "Epoch 371/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12259.1514 - r_square: -0.1477 - val_loss: 6623.0928 - val_r_square: 0.1707\n",
      "Epoch 372/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 11605.8516 - r_square: -0.0866 - val_loss: 7457.4697 - val_r_square: 0.0662\n",
      "Epoch 373/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11638.2266 - r_square: -0.0896 - val_loss: 10542.9805 - val_r_square: -0.3202\n",
      "Epoch 374/400\n",
      "8/8 [==============================] - 0s 50ms/step - loss: 12722.5654 - r_square: -0.1911 - val_loss: 9718.9678 - val_r_square: -0.2170\n",
      "Epoch 375/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12150.0449 - r_square: -0.1375 - val_loss: 8543.4316 - val_r_square: -0.0698\n",
      "Epoch 376/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 12246.7510 - r_square: -0.1466 - val_loss: 6830.6250 - val_r_square: 0.1447\n",
      "Epoch 377/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11604.5869 - r_square: -0.0864 - val_loss: 6926.9067 - val_r_square: 0.1326\n",
      "Epoch 378/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 12099.5859 - r_square: -0.1328 - val_loss: 7018.0737 - val_r_square: 0.1212\n",
      "Epoch 379/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11265.1143 - r_square: -0.0547 - val_loss: 6387.1260 - val_r_square: 0.2002\n",
      "Epoch 380/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12242.4336 - r_square: -0.1462 - val_loss: 6717.7827 - val_r_square: 0.1588\n",
      "Epoch 381/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12442.4551 - r_square: -0.1649 - val_loss: 9497.7588 - val_r_square: -0.1893\n",
      "Epoch 382/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 10968.3945 - r_square: -0.0269 - val_loss: 7163.6729 - val_r_square: 0.1030\n",
      "Epoch 383/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 11532.4678 - r_square: -0.0797 - val_loss: 7764.2358 - val_r_square: 0.0278\n",
      "Epoch 384/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12185.0488 - r_square: -0.1408 - val_loss: 8867.5186 - val_r_square: -0.1104\n",
      "Epoch 385/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12429.6045 - r_square: -0.1637 - val_loss: 8759.7080 - val_r_square: -0.0969\n",
      "Epoch 386/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11905.4854 - r_square: -0.1146 - val_loss: 8984.0908 - val_r_square: -0.1250\n",
      "Epoch 387/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11555.7158 - r_square: -0.0819 - val_loss: 7599.9585 - val_r_square: 0.0483\n",
      "Epoch 388/400\n",
      "8/8 [==============================] - 0s 51ms/step - loss: 11866.1230 - r_square: -0.1109 - val_loss: 7569.3770 - val_r_square: 0.0522\n",
      "Epoch 389/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12802.1562 - r_square: -0.1986 - val_loss: 7377.5981 - val_r_square: 0.0762\n",
      "Epoch 390/400\n",
      "8/8 [==============================] - 0s 40ms/step - loss: 12804.5967 - r_square: -0.1988 - val_loss: 6912.8984 - val_r_square: 0.1344\n",
      "Epoch 391/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12782.7266 - r_square: -0.1967 - val_loss: 7082.7568 - val_r_square: 0.1131\n",
      "Epoch 392/400\n",
      "8/8 [==============================] - 0s 47ms/step - loss: 12618.0322 - r_square: -0.1813 - val_loss: 7720.0415 - val_r_square: 0.0333\n",
      "Epoch 393/400\n",
      "8/8 [==============================] - 0s 45ms/step - loss: 11846.0967 - r_square: -0.1090 - val_loss: 8051.9326 - val_r_square: -0.0082\n",
      "Epoch 394/400\n",
      "8/8 [==============================] - 0s 49ms/step - loss: 11426.1787 - r_square: -0.0697 - val_loss: 6743.0957 - val_r_square: 0.1556\n",
      "Epoch 395/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12541.3984 - r_square: -0.1741 - val_loss: 7495.2949 - val_r_square: 0.0615\n",
      "Epoch 396/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12667.8154 - r_square: -0.1860 - val_loss: 7424.9717 - val_r_square: 0.0703\n",
      "Epoch 397/400\n",
      "8/8 [==============================] - 0s 42ms/step - loss: 12526.8779 - r_square: -0.1728 - val_loss: 7228.3193 - val_r_square: 0.0949\n",
      "Epoch 398/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 11696.0264 - r_square: -0.0950 - val_loss: 7055.9248 - val_r_square: 0.1165\n",
      "Epoch 399/400\n",
      "8/8 [==============================] - 0s 46ms/step - loss: 12021.4355 - r_square: -0.1255 - val_loss: 6949.5635 - val_r_square: 0.1298\n",
      "Epoch 400/400\n",
      "8/8 [==============================] - 0s 43ms/step - loss: 12477.2568 - r_square: -0.1681 - val_loss: 7073.9824 - val_r_square: 0.1142\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.layers import Dense, Dropout, GRU, BatchNormalization\n",
    "\n",
    "model_gru2 = Sequential()\n",
    "\n",
    "model_gru2.add(GRU(128, input_shape=(10, 5), return_sequences=True))\n",
    "model_gru2.add(BatchNormalization())\n",
    "model_gru2.add(Dropout(0.2))\n",
    "model_gru2.add(GRU(64, return_sequences=True))\n",
    "model_gru2.add(BatchNormalization())\n",
    "model_gru2.add(Dropout(0.2))\n",
    "model_gru2.add(GRU(32))\n",
    "model_gru2.add(BatchNormalization())\n",
    "model_gru2.add(Dropout(0.2))\n",
    "model_gru2.add(Dense(1, activation='linear'))\n",
    "\n",
    "model_gru2.compile(loss='mean_squared_error', optimizer=tf.keras.optimizers.Adam(learning_rate=0.1), metrics=[tfa.metrics.RSquare()])\n",
    "\n",
    "\n",
    "history_gru2 = model_gru2.fit(X_train_rs, y_train1, epochs=400, batch_size=128, validation_data=(X_val_rs, y_val1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "id": "e486c371-3ccd-47d8-a05e-9db070da0ec1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 0s 8ms/step\n",
      "93.43407170017078\n"
     ]
    }
   ],
   "source": [
    "y_pred_gru2_tr = model_gru2.predict(X_train_rs)\n",
    "gru2_rmse_train = np.sqrt(mean_squared_error(y_train1, y_pred_gru2_tr))\n",
    "print(gru2_rmse_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6fed363b-8759-4204-8b00-dbca29d0c571",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}