|
|||||
|
|
Gene subset selection in kernel-induced feature space |
|
| Authors: | Satoshi Niijima Satoru Kuhara |
| Affiliation: | aDepartment of Bioinformatics, Graduate School of Systems Life Sciences, Kyushu University, Hakozaki 6-10-1, Higashi-ku, Fukuoka 812-8581, Japan bFaculty of Agriculture, Kyushu University, Hakozaki 6-10-1, Higashi-ku, Fukuoka 812-8581, Japan |
| Abstract: | This paper proposes a new filter approach to gene subset selection for kernel-based classifiers. We derive kernel forms of several well-known class separability criteria, and gene subset selection based on the kernelized criteria is applied to microarray cancer classification problems. The performance of our proposed strategy is compared in experiments with those of the conventional filter approach as well as gene ranking methods. |
| Keywords: | Microarray Gene selection Cancer classification Kernel methods |
| 本文献已被 ScienceDirect 等数据库收录！ | |