Multigranular rough set model based on robust intuitionistic fuzzy covering with application to feature selection

Abstract

Fuzzy and intuitionistic fuzzy β covering has attracted the interest of many researchers recently. However, some of the factors namely 1. the lack of inclusion relationship between lower and upper approximation, 2. Inability to fit real valued dataset i.e. misclassification, hinder its application in decision making. Also, majority of these approaches are affected by noise. All these factors have created the need for a robust model. This paper proposes a novel intutitionistic fuzzy β covering model which is robust, combining intuitionistic fuzzy set, β covering and multigranulation rough sets. A new intuitionistic fuzzy β covering based multigranulation model is proposed which satisfies inclusion relationship between two approximations. This robust model is used to formulate dependency degree, which is evaluated at various granularity levels and is thereby employed for feature subset selection. Series of experiments are conducted on real valued datasets to illustrate the robustness and effectiveness of the underlying model. Furthermore, a comparative analysis with state of art approaches followed by statistical test lays down the superiority of the proposed model. © 2023 Elsevier Inc.