Comparison study of different Feature selection methods for Cancer Subtype classification
Differential diagnosis of Cancer sub-types and thus their treatments are one of the most challenging problems in clinical Medicine.. To choose the right treatment plan, a proper diagnosis is to be made.This project is intended to accurately classify a patient’s cancer type based on gene expressions. We have modelled a SVM classifier using different feature selection methods (Variance threshold,Selectkbest,selectmodel,RFE and SFS) to find the top features that would help in the accurate prediction of Cancer subclasses.The refined data set with quality samples makes it easier for differential diagnosis