Tuning fuzzy rules based on fuzzy clustering and neuro-fuzzy methods

Based on fuzzy c-means clustering algorithms, we developed a neuro-fuzzy learning approach for tuning fuzzy rules under fuzzy singleton-type reasoning method. In the approach, we can design roughly initial tuning parameters before the learning which is regarded to be a difficult problem in the neuro-fuzzy processing, in general. Therefore, the fuzzy rules generated by using the learning algorithm are more than reasonable and satisfiable to the identified system model. Moreover, we shown the efficiency of the developed method by means of a numerical example.