An improved method of fuzzy time series model

The study of fuzzy time series has increasingly attracted much attention due to its salient capabilities of tackling uncertainty and vagueness inherent in data collected. However, two shortcomings of the existing fuzzy time series forecasting methods are that they lack persuasiveness in partitioning interval and fuzzifying data. This paper introduces a new fuzzy time series method based on fuzzy c-means (FCM) clustering. Firstly, a formula based on distance is proposed to calculate cluster number. Secondly, based on the cluster number, unequal-sized intervals are obtained. Thirdly, a new definition method of the fuzzy sets is objectively given by distance in data fuzzification. Finally, the optimal forecasting results are obtained by tuning the distance parameter and utilizing the standard error (RMSE). Meanwhile, the optimal cluster number is determined by the smallest standard error (RMSE). The forecasting of Alabama university enrollments shows that the method outperforms the existing some methods.