An Alternative Method of Backward Fuzzy Interpolation based on Areas of Fuzzy Sets

Abstract

Fuzzy rule interpolation techniques can reduce the complexity of fuzzy systems and make inferences of conclusions in sparse rule-based systems. However, when certain crucial antecedents are missing in the observations and the subsequent interpolation inference process involves these missing antecedents, conventional fuzzy interpolation methods cannot obtain inferred conclusions. To tackle this problem, Jin et al. proposed the method of backward fuzzy rule interpolation, which allows the missing antecedents to be deduced or interpolated from the known antecedents and given conclusion, extending the research field of fuzzy interpolation techniques. In order to extend the generality of backward fuzzy rule interpolation inference, this paper proposes a backward fuzzy rule interpolation approach based on the CCL algorithm, which utilizes the triangular fuzzy membership functions and verifies the effectiveness of the approach by examples.