Learning method of fuzzy functional inference model

Seki and Mizumoto have proposed a fuzzy functional inference method in which the consequent parts of the T-S inference method is generalized to fuzzy functions, However, it is noted that the adjustment of fuzzy functions is generally difficult. Therefore, in this paper, we propose a learning method of the fuzzy functional inference method by using the base of the fuzzy functions. Moreover, the fuzzy functional inference method by using the proposed learning method is shown to be superior to the conventional method by applying to a medical diagnosis as an example of application to real systems.