Online MOACO biclustering of microarray data

Multi-objective optimization (MOP) a fast growing area of research. Bioinformatics data sets come mostly from DNA microarray experiments. The analysis of microarray data sets can provide valuable information on the biological relevance of genes and correlations among them. Biclustering methods allow us to identify genes with similar behavior with respect to different conditions. A single bicluster represents a given subset of genes in a given subset of conditions. For solving multiple objectives optimization, ant colony optimization algorithms have been shown to be very effective for MOP. This paper proposes online Multiple Objective Ant Colony Optimization biclustering algorithm to solve patterns mining problem of microarray dataset. During optimization, the size of ant population is dynamically changed to quicken the convergence of the algorithm. Experimental analysis on two real dataset shows that the proposed algorithm achieves good performance in the diversity of solution and the time complexity of the algorithm.