Gene selection in cancer classification using hybrid method based on Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC) feature selection and support vector machine

This study proposed the Hybrid method, Particle Swarm Optimization-Support Vector Machine (PSO-SVM) and Artificial Bee Colony-Support Vector Machine (ABC–SVM), in selecting informative genes for cancer classification. PSO and ABC are filter methods for eliminating inefficient genes in high-dimensional gene expression data using ranking techniques. Top ranking genes are chosen as informative genes. While SVM is used to eliminate excessive genes after being filtered by PSO and ABC, it can produce more accurate gene expression data. The informative genes chosen by PSO-SVM and ABC-SVM will be used for cancer classification. Among the two methods, ABC-SVM is the best method in classifying cancer with an accuracy rate of 88 %. All these datasets were obtained from UCI Machine Learning Repository.