Sequential Adaptive Fuzzy Inference System (SAFIS) for nonlinear system identification and prediction

In this paper, 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 (FIS). 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 do not warrant adding of fuzzy rules, then only the parameters of the “closest” (in a Euclidean sense) rule are updated using an extended kalman filter (EKF) scheme. The performance of SAFIS is compared with several existing algorithms on two nonlinear system identification benchmark problems and a chaotic time series prediction problem. Results indicate that SAFIS produces similar or better accuracies with less number of rules compared to other algorithms.