Microarray data classification based on evolutionary multiple classifier system

Conference Name:2011 3rd International Conference on Mechanical and Electronics Engineering, ICMEE 2011. Conference Address: Hefei, China. Time:September 23, 2011 - September 25, 2011.Hefei University of TechnologyDesigning 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. 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. When this GA stops, a set of base classifiers are generated. Here we use all the nondominated individuals in last generation to build an ensemble system and test the proposed ensemble method and the method that apply a classifier selection process to select proper classifiers from all the individuals in last generation. The experimental results show the proposed ensemble method is roubust and can lead to promising results. 漏 (2012) Trans Tech Publications, Switzerland