Gene selection in cancer classification using PSO/SVM and GA/SVM hybrid algorithms

Abstract

Abstract-In this work we compare the use of a Particle Swarm Optimization (PSO) and a Genetic Algorithm (GA) (both augmented with Support Vector Machines SVM) for the classification of high dimensional Microarray Data. Both algorithms are used for finding small samples of informative genes amongst thousands of them. A SVM classifier with 10-fold cross-validation is applied in order to validate and evaluate the provided solutions. A first contribution is to prove that P SOSV M is able to find interesting genes and to provide classification competitive performance. Specifically, a new version of PSO, called Geometric PSO, is empirically evaluated for the first time in this work using a binary representation in Hamming space. In this sense, a comparison of this approach with a new GASV M and also with other existing methods of literature is provided. A second important contribution consists in the actual discovery of new and challenging results on six public datasets identifying significant in the development of a variety of cancers (leukemia, breast, colon, ovarian, prostate, and lung).

The Performance of Bio-Inspired Evolutionary Gene Selection Methods for Cancer Classification Using Microarray Dataset

“…In other hand, compared with other approach Bio-Inspired evolutionary approach generally obtains one gene subset with better classification performance but much more computational cost [3]. Bio-Inspired evolutionary gene selection methods can be impractical for some computationally expensive algorithms such as SVM or artificial neural networks [2]. However, the classification accuracy in cancer research in significant, because this reason often uses Bio-Inspired evolutionary gene selection methods in cancer prediction [3].…”

“…It offers an efficient method of gathering data that can be used to determine the patterns of gene expression of all the genes in an organism in a single experiment [1], [2]. DNA microarrays can be used to determine which genes are being expressed in a given cell type at a particular time and under particular conditions, to compare the gene expression in two different cell types or tissue samples, and then we can determine the more informative genes that are responsible to cause specific disease or cancer [3].…”

A Comparative Study on Bioinformatics Feature Selection and Classification

“…For this reason, tools for decreasing the number of features in order to improve the classification or to help to identify interesting features in noisy environments are necessary. In addition, SVM can treat data with a large number of features, but it has been shown that its performance is increased by reducing the number of features [7].…”

“…For non linearly separable cases, samples are mapped to a high dimensional space where such a separating hyperplane can be found. The assignment is carried out by means of a mechanism called the kernel function [7]. SVM is widely used in the domain of cancer studies, protein identification and specially in Microarray data.…”

A hierarchical two-phase framework for selecting genes in cancer datasets with a neuro-fuzzy system

“…It needs to use computational methods in that gene expression microarray datasets are very large number of various genes in the gene datasets. It has been a challenging problem to identify the genes that are relevant or not to a clinical diagnosis [1,5,15]. As feature selection methods, mutual information [6,7,13], the t-test [14], threshold number of misclassifications (TNoM) score [8], and the Bhattacharyya distance [3,4,21] have been widely used in finding relevant genes.…”