A novel approach to select important genes from microarray data

Feature subset selection is a well-known pattern recognition problem, which aims to reduce the number of features used in classification or recognition. This reduction is expected to improve the performance of classification algorithms in terms of speed, accuracy and simplicity. Most existing feature selection investigations are not suitable for microarray data, so this paper focuses on gene selection problem. The main contributions of this paper are that a new feature selection method A-score is introduced and constructed an improved fuzzy Bayesian classifier. We evaluate the performance of A-score using three well-known benchmark data sets: the iris data, the wine data, and the Wisconsin breast cancer data and two microarray data: ALL-AML Leukemia and colon cancer. In general, A-score can significantly reduce the number of genes, and perform better than T-score and C-score.