RFEGSA: Gene Expression Feature Selection Utilizing Recursive Feature Elimination and Gravitational Search Algorithm

Gene expression data usually present the characteristics of high dimension and small sample size. In such data, it is crucial to conduct feature selection to reduce dimensions and retain key information. However, accurately identifying important features remains challenging, and it is of great significance for precision medicine and drug development. To solve this problem, this paper proposes a three-stage feature selection method. First, this study integrated three filter algorithms to screen out the features with higher rankings. Then, the feature subset with the optimal prediction accuracy is determined through the recursive feature elimination (RFE) algorithm. Finally, the gravitational search algorithm (GSA) is utilized to eliminate redundant features. Experiments on the gene expression dataset show that the performance of this method is superior to that of 7 popular feature selection algorithms.