Use of multiobjective differential fuzzy clustering with ANN classifier for unsupervised pattern classification: Application to microarray analysis

Microarray technology has made it possible to monitor the expression levels of many genes simultaneously across a number of experimental conditions. Fuzzy clustering is an important tool for analyzing microarray gene expression data. However, this fact motivated us to present a new multiobjective evolutionary algorithm using differential evolution and artificial neural network (ANN) classifier for fuzzy clustering. The proposed approach has two-phases, in one phase, it optimizes multiple cluster validity measures simultaneously to get the resultant set of near-pareto-optimal solutions which contains number of nondominated solutions and in other phase, fraction of the data points selected from different clusters based on their proximity to the respective centres of each clusters for each of the nondominated solutions and thereafter fuzzy voting technique is used to generate the train data set for ANN classifier to classify the remaining data points. Our hybrid approach is tested for two publicly available benchmark microarray data sets. Our results are compared with respect to the multiobjective genetic fuzzy clustering algorithm that used NSGA-II, which is a state-of-the-art for MOEA. Also biological significance test has been carried out using a web based gene annotation tool to show that the proposed method is able to produce biologically relevant clusters of co-expressed genes.