based on the Universal Approximation Theory of Fuzzy system
presents a fuzzy traffic predictor using nearest neighborhood clustering learning algorithm.Its prediction effect is examined through autoregressive (AR) process and real video traffic.The results show that the approximator/predictor is accurate and flexible.At last
the applications of the predictor in dynamic bandwidth allocation is discussed.