NEURO-FUZZY ALGORITHM FOR THE SYNTHESIS OF FUZZY INFERENCESYSTEMS

Аннотация

A neuro-fuzzy algorithm for the synthesis of fuzzy inference systems is proposed. A two-stage adaptive algorithm for FIS synthesis is described. In the first stage, clustering of initial fuzzy parameters is performed in order to reduce the number of input parameters in fuzzy rules, and in the second stage, the synthesis of fuzzy models (inference rules) of the Sugeno type is carried out.