Clustering in conjunction with quantum genetic algorithm for relevant genes selection for cancer microarray data

Abstract

Quantum Genetic Algorithm, which utilizes the principle of quantum computing and genetic operators, allows efficient exploration and exploitation of large search space simultaneously. It has been used recently to determine a reduced set of features for cancer microarray data to improve the performance of the learning system. However, the length of the chromosome used is the original dimension of the feature vector. Hence, despite the use of the quantum variant of GA, it requires huge memory and computation time for high dimensional data like microarrays. In this paper, we propose a two phase approach, ClusterQGA, that determines a minimal set of relevant and noniredundant genes. Experimental results on publicly available cancer microarray datasets demonstrate the effectiveness of the proposed approach in comparison to existing methods in terms of classification accuracy and number of features. Also, the proposed approach takes less computation time in comparison to Genetic quantum algorithm proposed by Abderrahim et al. © Springer-Verlag 2013.