Information Fusion Technique for Fuzzy Time Series Model

This paper proposes a high order multiple attribute fuzzy time series model, which corporate a clustering method and order weighted averaging (OWA) operator. The proposed model can deal with (1) lacking persuasiveness in determining the universe of discourse and the linguistic length of intervals, and (2) only one attribute is usually considered in forecasting not multiple attributes. Furthermore, we can according to different situation α to adjust high order forecasting value. To verify the proposed model, we use the yearly data on enrollments at the University of Alabama and TAIFEX (Taiwan Futures Exchange) as experimental datasets. Finally, this paper compares forecasting performances of proposed methods with Hang et al.’s [1] and Cheng et al.’s [2] models, the results of empirical analysis conclude that the proposed model surpasses in accuracy the listing models.