Ant colony optimisation classification for gene expression data analysis

Abstract

Microarray studies and gene expression analysis have received much attention over the last few years and provide promising avenues towards the understanding of fundamental questions in biology and medicine. In this paper we investigate the application of ant colony optimisation (ACO) based classification for the analysis of gene expression data. We employ cAnt-Miner, a variation of the classical Ant-Miner classifier, to interpret numerical gene expression data. Experimental results on well-known gene expression datasets show that the ant-based approach is capable of extracting a compact rule base and provides good classification performance. © 2009 Springer-Verlag Berlin Heidelberg.