A Hybrid Fuzzy Time Series Forecasting Model
Dong Shu-l · Mohu xitong yu shuxue · 2014
With the needs of theory and application,the research and application of fuzzy time series model has been widely studied.This paper improves the traditional fuzzy time series model from aspects of partition of interval and extraction of fuzzy rules.Firstly,the model adopts the method of automatic clustering to divide universe of discourse,and then establishs fuzzy rules with weights.Secondly,particle swarm algorithm is used to further improve the prediction accuracy.Finally,the enrollment of Alabama University is used as the experimental data for the forecasting models.The experimental results show that the proposed model outperforms the compared models.