A Novel SVM-RFE for Gene Selection

摘要：

Selecting a subset of informative genes from microarray expression data is a critical data preparation step in cancer classification and other biological function analysis. The support vector machine recursive feature elimination (SVM-RFE) is one of the most effective feature selection method which has been successfully used in selecting informative genes for cancer classification. While, the SVM-RFE selects genes only using the gene expression data without using any other biological information of the genes. Based on the biology information of the genes, it may be beneficial to identify the genes that are relevant to the cancer. We propose a novel SVM-RFE method for gene selection by incorporating the Kyoto Encyclopedia of genes and genomes (KEGG) pathway information into feature selection process. Numerical results indicate that the novel SVM-RFE tends to provide better variable selection results than the SVM-RFE.

关键词： Support vector machine microarray data gene selection

作者: Jun-Yan Tan Zhi-Xia Yang Naiyang Deng

作者单位: College of Science,China Agricultural University,100083,Beijing,ER.China College of Mathematics and Systems Science,Xinjiang University,830046,Urumuqi,P.R.China Academy of M

会议类型: 国际会议

会议名称: The 3rd International Symposium on Optimization and System Biology(第三届最优化与系统生物学国际会议 OSB09)

会议地点: 张家界

会议语种:英文

页码: 237-244

在线出版日期: 2009-09-20（万方平台首次上网日期，不代表论文的发表时间）