Fuzzy Model Identification Method Based on Tikhonov Regularization

We consider the ill-posedness of the fuzzy system identification process. The standard fuzzy c-means clustering algorithm is used to divide the input space, and fuzzy rules are extracted from the known input data in the system. To counteract the ill-posedness in the consequent parameter identification process, we apply the Tikhonov regularization method and introduce the regularized functional in the minimizing functional to solve ill-posed problems. Then we use the Bayesian method to calculate the regularization parameter, and we give the specific algorithm. Simulation results show that this method has well-posedness.