時系列データのファジィ化と逐次更新モデル

抄録

A fuzzy autoregressive (AR) model proposed by Yabuuchi are an interval-type AR model. His fuzzy AR model uses fuzzy time-series data. As computations based on fuzzy numbers increase vagueness, an intervaltype AR model may produce unnatural predicted values. This paper solved this problem using a numerical example by two approaches. The first is to change the target of fuzzification from time-series data to autocovariances. The second is to build the interval-type autoregressive model sequentially. As the two approaches solved the problem, this paper considers the results used two approaches.