Clustering of gene expression data with quantum-behaved particle swarm optimization

Abstract

Based on the previously proposed Quantum-behaved Particle Swarm Optimization (QPSO) algorithm, in this paper, we focus on the application of QPSO in gene expression data clustering which can be reduced to an optimization problem. The proposed clustering algorithm partitions the N patterns of the gene expression dataset into user-defined K categories to minimize the fitness function of Total Within-Cluster Variation. Thus a partition with high performance is obtained. The experiment results on four gene expression data sets show that our QPSO-based clustering algorithm will be an effective and promising tool for gene expression data analysis. © 2008 Springer-Verlag Berlin Heidelberg.