Utility of Binary Cuckoo Search Approach for Microarray Data Analysis

Abstract

In the recent literature, many metaheuristic approaches has been developed to examine, interpret, and analyze high dimensional data. However, there is always a requirement to design a more productive and cost-effective technique. Many of the authors have acknowledged the capability of cuckoo search approach. Inspired from the blooming application of cuckoo search algorithm (CSA), the authors have proposed a binary cuckoo search algorithm (BCSA) for gene selection and classification of microarray data. With the objective of optimizing gene subset and classification result, BCSA has been implemented. BCSA is optimized using support vector machine (SVM) classifier, and significant gene set is extracted. This experimental study has been conducted over five microarray datasets. Superiority of this gene subset is shown using a few other classifiers such as k-nearest neighbor and artificial neural network. The model performance is also compared with a few other metaheuristic approaches such as particle swarm optimization, genetic algorithm, artificial bee colony algorithm, and differential evolution.