Vis enkel innførsel

dc.contributor.advisorLiland, Kristian Hovde
dc.contributor.advisorTomic, Oliver
dc.contributor.authorSana, Hemanth Babu
dc.date.accessioned2021-10-11T11:09:20Z
dc.date.available2021-10-11T11:09:20Z
dc.date.issued2021
dc.identifier.urihttps://hdl.handle.net/11250/2789013
dc.description.abstractGastrointestinal carcinoma are the cancers that affect the gastrointestinal tract and other organs that include esophagus, pancreas, stomach, colon, rectum, anus, liver and intestine. Gastrointestinal cancers account to 26% of global cancer incidence. They account to 35% of all cancer-related deaths. Being able to find the factors responsible for increasing the life span of patients adds significant value in the course of treatment for the doctors. This Master’s thesis explored the feasibility of employing two new state-of-art techniques- Repeated Elastic Net Technique (RENT) for feature selection and Sequential and Orthogonalized Partial Least Squares regression (SO-PLS). This study helped to (1) find features that are important for predicting the target using RENT and to (2) use the underlying dimensionality of the data blocks to explain the variance of the target using SO-PLS. The feature selection using RENT proved to be useful by reducing the number of features from 57 to 7 in the first block and from 27 to 7 in the second block. By using these selected features from both blocks, SO-PLS regression achieved a cumulative calibrated explained variance of 76.4%. The score and loading plots from SO-PLS helped in identifying the features that explain the distribution of values in the target block. These results indicate that RENT and SO-PLS have the potential in developing as useful techniques for clinicians in understanding the factors responsible for the longevity of patient life.en_US
dc.language.isoengen_US
dc.publisherNorwegian University of Life Sciences, Åsen_US
dc.rightsAttribution-NonCommercial-NoDerivatives 4.0 Internasjonal*
dc.rights.urihttp://creativecommons.org/licenses/by-nc-nd/4.0/deed.no*
dc.subjectSO-PLSen_US
dc.subjectRENTen_US
dc.subjectPLS regressionen_US
dc.subjectElastic Neten_US
dc.subjectmultiblock analysisen_US
dc.titleSequential and orthogonalized partial least squares regression applied to healthcare data acquired from patients diagnosed with gastrointestinal carcinomaen_US
dc.typeMaster thesisen_US
dc.subject.nsiVDP::Technology: 500en_US
dc.description.localcodeM-DVen_US


Tilhørende fil(er)

Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel

Attribution-NonCommercial-NoDerivatives 4.0 Internasjonal
Med mindre annet er angitt, så er denne innførselen lisensiert som Attribution-NonCommercial-NoDerivatives 4.0 Internasjonal