A Novel Feature Selection Algorithm using Particle Swarm Optimization for Cancer Microarray Data

Microarray data are often extremely asymmetric in dimensionality, highly redundant and noisy. Most genes are believed to be uninformative with respect to studied classes. This paper proposed a novel feature selection approach for the classification of high dimensional cancer microarray data, which used filtering technique such as signal-tonoise ratio (SNR) score and optimization technique as Particle swarm Optimization (PSO). The proposed method is divided in to two stages. In the first stage the data set is clustered using k-means clustering, SNR score is used to rank each gene in every cluster. The top scored genes from each cluster is gathered and a new feature subset is generated. In the second stage the new feature subset is used as input to the PSO and optimized feature subset is being produced. Support vector machine (SVM), k-nearest neighbor (k-NN) and Probabilistic Neural Network (PNN) are used as evaluators and leave one out cross validation approach is used for validation. We have compared both of our approach and approaches using PSO in the literature. It has been demonstrated that our approach using PSO gives better result than others.