Feature Extraction for Cancer Classification Using Kernel-Based Methods

Abstract

In this paper, kernel-based feature extraction method from gene expression data is proposed for cancer classification. The performances of four kernel algorithms, namely, kernel Fisher discriminant analysis (KFDA), kernel principal component analysis (KPCA), kernel partial least squares (KPLS), and kernel independent component analysis (KICA), are compared on three benchmarked datasets: breast cancer, leukemia and colon cancer. Experimental results show that the proposed kernel-based feature extraction methods work well for three benchmark gene dataset. Overall, the KPLS and KFDA show the best performance, and KPCA and KICA follow them.KeywordsSupport Vector MachineKernel MatrixCancer DatasetKernel Principal Component AnalysisCancer ClassificationThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.