On the equivalence of a RBF-like network to TS fuzzy systems: a GA approach for TS-network

Abstract

Functional equivalence of radial basis function (RBF) networks and a class of fuzzy inference systems is considered. The class of fuzzy systems based on the Takagi-Sugeno model is referred to as TS-model of fuzzy inference. From the abstract mathematical viewpoint the functional equivalence between radial basis function networks and fuzzy inference systems is already shown. However, from the viewpoint of realisation of the models with data, the difference between the model performances is observed. The research makes comparisons between an RBF model and its fuzzy model counterpart and qualifications on the equivalence being observed are reported.