Biclustering gene expression data using KMeans-binary PSO hybrid

Biclustering is a very useful data mining technique which identifies coherent patterns from microarray gene expression data. A bicluster of a gene expression dataset is a subset of genes which exhibit similar expression patterns along a subset of conditions. Biclustering is a powerful analytical tool for the biologist and has generated considerable interest over the past few decades. The problem of locating the most significant biclusters in gene expression data has shown to be NP complete. In this paper a PSO based algorithm is developed for biclustering gene expression data. This algorithm has three steps. In the first step high quality bicluster seeds are generated using KMeans clustering algorithm. From these seeds biclusters are generated using particle swarm optimization. In the third stage an iterative search is performed to check the possibility of adding more genes and conditions within the given threshold value of mean squared residue score. Experimental results on real datasets show that our approach can effectively find high quality biclusters.