Feature Selection for Cancer Classification Based on SRPSO Algorithm

Abstract

Improving the accuracy of cancer classification plays an important role in cancer-assisted diagnosis. Genes selection is an important factor for improving the accuracy of cancer classification. In this paper, based on the standard particle swarm optimization algorithm, an SRPSO algorithm with self-adaptive and reverse-learning mechanism is proposed. It is applied to select feature genes from microarray datasets, and the results are used for cancer classification via SVM to make 5-fold cross-validation. To evaluate the performance of SRPSO, four different cancer datasets including Colon, ALL_AML, MLL, and SRBCT were selected. Based on the evaluation process, the SRPSO algorithm provided better results on each dataset.