A Feature Selection Method on Gene Expression Microarray Data for Cancer Classification Abstract

In medical data extraction, the gene dimension is often much larger than the sample size. To address this issue, we need to use a feature selection algorithm to select gene feature subsets with a strong correlation with the phenotype to ensure the accuracy of subsequent analyses. This research presents a new three-stage hybrid gene feature selection method, which combines a variance filter, extremely randomized tree, and whale optimization algorithm. Initially, a variance filter is employed to reduce the dimension of the gene feature space, and then an extremely randomized tree is utilized to further reduce the gene feature set. Finally, the whale optimization algorithm is applied to select the optimal gene feature subset. We evaluated the proposed method using the K-nearest neighbors (KNN) classifier on four published gene expression profile datasets and compared it with other gene selection algorithms. The results demonstrate that the proposed method has significant advantages in various evaluation indicators.