A genetic algorithm applied to optimal gene subset selection

Abstract

Optimal gene subset selection plays an important role in classification of patient samples. Different to existed methods, we propose a novel optimal gene subset selection approach based on genetic algorithms (GAs). Special fitness function is applied in this scheme. Going beyond other methods, this GA-based method automatically determines the members of a predictive gene group, as well as the optimal group size. The evaluation experiments are applied to two data sets. The results and some discussions are presented too.