Sequential Adaptive Fuzzy Inference System

In this chapter, a Sequential Adaptive Fuzzy Inference System called SAFIS is developed based on the functional equivalence between a radial basis function network and a fuzzy inference system. In SAFIS, the concept of “Influence” of a fuzzy rule is introduced and using this the fuzzy rules are added or removed based on the input data received so far. If the input data does not warrant adding of fuzzy rules, then only the parameters of the “closest” (in a Euclidean sense) rule are updated using EKF scheme. Moreover, improvements to SAFIS for enhancing its performance in both accuracy and speed are described in the chapter and the resulting algorithm is referred to as extended SAFIS (ESAFIS). Different from the SAFIS where the zero-order TS model is utilized, the developed ESAFIS algorithm suits both the zero-order and first-order TS fuzzy models. Empirical study of SAFIS and ESAFIS with the two fuzzy models is executed based on several commonly used classification benchmark problems. The results indicate that the proposed ESAFIS produces higher classification accuracy with reduced computational complexity compared with SAFIS and other algorithms.