Covering rough set-based three-way decision feature selection

Abstract

Feature selection is the process of selecting a subset of features from the entire dataset such that the selected subset can be used on behalf of the entire features of dataset to reduce the complexity of further processing. Recently, many feature selection approaches have been developed effectively based on rough sets. To provide a reasonable semantic interpretation for the three regions of probability rough sets and decision theoretic rough sets, the three-way decision theory is proposed. By using of the covering rough set model, the positive region, boundary region and negative region of the three-way decision approach are established to conduct feature selection. In this paper, the covering rough set-based three-way decision feature selection method (CTFS) is proposed to extend the traditional rough set version. Furthermore, the condition entropy based on three-way decision regions can be used as an evaluation method to select features. The experimental results show that the proposed feature selection approach results in a smaller selected feature subsets compared to the state-of-the-art feature selection methods with higher classification accuracies.