A study on automatic allocation of membership functions for fuzzy modeling

Fuzzy modeling is a method to describe input-output relationships using fuzzy inference. Fuzzy inference is capable of approximating nonlinear systems. The paper proposes a method of fuzzy modeling for automatic allocation of membership functions. The initial allocation of the membership function is searched by roughly allocating them under the Akaike information criterion. The inserting points of membership functions are finely searched by the new method. Experiments are done to show that the proposed method is effective in allocating the membership functions more properly than the conventional method