Prediction of Tumor Location in Prostate Cancer Tissue using Gene Expression

Abstract

prostate cancer can be missed due to the limited number of biopsies or the ineffectiveness of standard screening methods. In this work, a classification model is built based on gene expression measurements of samples from patients who have cancer on the left, right, and both lobes of the prostate as classes. A hyper feature selection is used to select the best possible set of genes that can differentiate the three classes. Standard machine learning classifiers with the one-versus-all technique are used to select literality biomarkers for each class.