关键词: 基因选择, 样本不平衡, 0-1规划, 分类

Abstract: Gene selection is one of important problems in gene expression data analysis.Although several gene selection methods have been proposed，yet there is no method simultaneously considering the problem of sample imbalance and the interaction of genes.However，the sizes of sample classes in microarray data are often unbalanced.Referring to cluster validation index，this paper proposes the 0-1 programming model of gene selection to answer for the problem of sample imbalance and gene interaction.Furthermore，a heuristic algorithm based on greedy strategy is proposed to solve the proposed optimization problem.The experimental results on three real microarray datasets show that the proposed model and algorithm are very efficient and robust to select discriminator genes.

Key words: gene selection, sample imbalance, 0-1 programming, classification