The design of evolutionary multiple classifier system for the classification of microarray data

Abstract

Designing an evolutionary multiple classifier system (MCS) is a relatively new research area. In this paper, we propose a genetic algorithm (GA) based MCS for microarray data classification. In detail, we construct a feature poll with different feature selection methods first, and then a multi-objective GA is applied to implement ensemble feature selection process so as to generate a set of classifiers. Then we construct an ensemble system with the individuals in last generation in two ways: using the nondominated individuals; using all the individuals accompanied with a classifier selection process based on another GA. We test the two proposed ensemble methods based on two microarray data sets, and the experimental results show that these two methods are robust and can lead to promising results. © 2011 Springer-Verlag.