Fuzzy Modeling by Fuzzy Clustering and Locally Weighted Regression

Abstract

A method of identification from input-output data is proposed for a fuzzy model in which the premise part of an if-then rule has a generalized form of the Takagi-Sugeno's fuzzy model This method consists of locally weighted regression for each observation and fuzzy clustering with respect to sample regression coefficients. Efficacy of the proposed method is demonstrated through numerical examples and an application to real data.