Multipolar Aggregation Operators in Reasoning Methods for Fuzzy Rule-Based Classification Systems

The reasoning methods in fuzzy rule-based classification systems are studied, and their relation to multipolar aggregation operators is discussed. We describe the change in the output of the classification system based on the change of the fuzzy reasoning method. We propose a new fuzzy reasoning method based on multipolar fusion OWA operator which can control the transition between the maximum rule method and the maximum vote method by a weighting vector. This new method uses principles of OWA operators and assigns the rule weight according to the strength of activation of the rule. We compare different fuzzy reasoning methods in case studies.