Adaptive Time-Variant Model Optimization for Fuzzy-Time-Series Forecasting

Fuzzy time series forecasting model is one of the tools that can be used to identify factors in order to solve the complex process and uncertainty, nowadays widely used in forecasting problems, but having appropriate universe of discourse and interval length are two subjects that exist in the Fuzzy time series. Recently Adaptive Time-Variant Model for fuzzy time series (ATVF) has been proposed with a computational method and an adaptive selection of analysis windows. In this paper, first we have introduced particle swarm optimization algorithm which is used for interval lengths improvement for ATVF model, another challenge that ATVF model confront with it is universe of discourse and this problem is solved using K-means clustering algorithm. Two models are applied to predict three data bases (the Enrolment of University of Alabama, Taiwan Futures Exchange (TAIFEX) and Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX)). The experimental results show that the proposed methods gets good forecasting results as compared to other existing fuzzy-time-series forecasting