NeuFuz: an intelligent combination of fuzzy logic with neural nets

Abstract

A novel method is presented to combine neural nets with fuzzy logic. The combined technology, NeuFuz, generates fuzzy logic rules and membership functions by learning the system behavior using input-output data. The generated rules and membership functions are then processed using new fuzzy logic algorithms for defuzzification, rule evaluation and antecedent processing which are developed based on neural network architecture and learning. These fuzzy logic algorithms replace conventional heuristic fuzzy logic algorithms and enable full mapping of neural net to fuzzy logic. Full mapping provides an important key feature of generating fuzzy rules and membership functions to meet a pre-specified accuracy level. NeuFuz also significantly improves performance, reliability, reduces design time and minimizes system cost by optimizing number of rules and membership functions.