Adaptive weighted fuzzy rule interpolation based on ranking values and similarity measures of rough-fuzzy sets

In real-world applications, the performances of traditional fuzzy interpolation methods are not good enough because they may produce inconsistent fuzzy interpolation results when contradictory fuzzy interpolated results are obtained after fuzzy interpolation processes. In recent years, several adaptive fuzzy rule interpolation (AFRI) methods have been proposed for identification and correction of the contradictory fuzzy interpolated results. In this paper, we propose a new adaptive weighted fuzzy rule interpolation (AWFRI) method based on ranking values and similarity measures of rough-fuzzy sets (RFSs). If the degree of similarity obtained between the fuzzy interpolation results is less than a predefined threshold value, the proposed AWFRI method applies the proposed transformation and modification techniques to solve the contradiction between fuzzy interpolation results. The experimental results show that the proposed AWFRI method can overcome the drawbacks of the existing AFRI methods because it obtains higher consistency degrees between fuzzy interpolation results and it deals with AWFRI based on RFSs rather than type-1 fuzzy sets or interval type-2 fuzzy sets. (C) 2019 Elsevier Inc. All rights reserved.