A new approach for calculating implications of fuzzy rules

Fuzzy inference systems have been pervasive in many areas of applications concerning machine learning and decision analysis where numerical computation and human heuristics integrate. This paper introduces an approach for computing the implications of fuzzy rules that consist of multiple input variables. Input variables are partitioned on the basis of information gain, and multiple secondary input attributes are computed by the Choquet fuzzy integral. The proposed model is computationally efficient in both capturing and aggregating attributes at detailed levels.