Heuristic optimization for Biclustering microarray gene Expression data

Abstract

Microarray data analysis is a critical step in practical applications newlineand often achieved with the help of two different data mining techniques such newlineas supervised classification and unsupervised classification clustering The newlinegoal of clustering is to extract information on how gene expression levels newlinevary among the different conditions In the context of clustering analysis newlinebiclustering is an interesting approach which aims to find the subsets of genes newlinethat have similar behaviour under subsets of conditions simultaneously newlineIn this thesis the general problem of biclustering is studied newlineVarious biclustering methods are considered and analysed of their merits and newlinedemerits The biclustering process is a complex one than clustering It is newlineproven to be a Nondeterministic Polynomialtime NPhard problem In newlineorder to overcome this issue heuristic algorithms are proposed It is used to newlinefind the coherent biclusters with minimum Mean Square Residue MSR and newlinelarge row variance in a large microarray gene expression data newline newline