Abstract
Microarray technology has provided biologists with the ability to measure the expression levels of thousands of genes in a single experiment. One of the urgent issues in the use of microarray data is the selection of a smaller subset of genes from the thousands of genes in the data that contributes to a disease. This selection process is difficult due to many irrelevant genes, noisy genes, and the availability of the small number of samples compared to the huge number of genes (higher-dimensional data). In this study, we propose a three-stage gene selection method to select a smaller subset of informative genes that is most relevant for the cancer classification. It has three stages: 1) pre-selecting genes using a filter method to produce a subset of genes; 2) optimising the gene subset using a multi-objective hybrid method to yield near-optimal gene subsets; 3) analysing the frequency of appearance of each gene in the different near-optimal gene subsets to produce a smaller subset of informative genes. The experimental results show that our proposed method is capable in selecting the smaller subset to obtain better classification accuracies than other related previous works as well as four methods experimented in this work. Additionally, a list of informative genes in the best gene subsets is also presented for biological usage.
