Radial basis function based adaptive fuzzy systems and their applications to system identification and prediction

In this paper we describe a neuro-fuzzy system with adaptive capability to extract fuzzy If-Then rules from input and output sample data through learning. The proposed system, called radial basis function (RBF) based adaptive fuzzy system (AFS), employs the Gaussian functions to represent the membership functions of the premise part of fuzzy rules. Three architectural deviations of the RBF based AFS are also presented according to different consequence types such as constant, first-order linear function, and fuzzy variable. These provide versatility of the network to handle arbitrary fuzzy inference schemes. We present examples of system identification and time series prediction to illustrate how to solve these problems and to demonstrate its validity and effectiveness using the RBF based AFS.