Fuzzy if-then rule generation based on neural network and clustering algorithm techniques

This paper proposes a self-tuning method for fuzzy if-then rule generation based on neural network and clustering algorithm techniques. In the tuning approach, the initial parameters of fuzzy rules are roughly designed by using a fuzzy clustering algorithm, and then fuzzy rules under fuzzy singleton-type reasoning are tuned by using a neuro-fuzzy learning algorithm. By this approach, the learning time can be reduced and the generated fuzzy rules are reasonable and suitable for the identified system. Finally, identifying a nonlinear function shows the efficiency of the tuning approach.