A fuzzy rule extraction method based on Dempster–Shafer theory

A fuzzy rule generation method which applies Dempster–Shafer theory (DST) to fuzzy inference systems is proposed. The sample data located in the same fuzzy input subspace are regarded as evidences, and all candidate rule consequents constitute the frame of discernment. Dempster’s combination rule is employed to fuse evidences to determine the rule consequent, which solves the problem of conflicting rules simultaneously. In addition, a BPA-based approach to optimize constant rule consequent is proposed. Experimental results illustrate that both the proposed rule extraction method and the optimization method for rule consequent have better performance.