Fuzzy forecasting based on high-order fuzzy logical relationships and automatic clustering techniques

Abstract

Fuzzy time series models have been widely used to handle forecasting problems, such as forecasting enrollments, temperature, and the stock index. If we can get better forecasting accuracy rates, then we can get more benefits. In this paper, we present a new method to handle forecasting problems using high-order fuzzy logical relationships and automatic clustering techniques. The proposed method uses the proposed automatic clustering algorithm to partition the universe of discourse into different lengths of intervals. We also apply the proposed method to forecast the enrollments of the University of Alabama, the temperature and the TAIFEX. The experimental results show that the proposed method gets a higher average forecasting accuracy rate than the existing methods.