Study on Fuzzy-Clustering Methods in Fuzzy Time Series Forecasting with Application

Abstract

In this research we study some fuzzy clustering methods in fuzzy time series forecasting especially Chen method and we used a FCM method to construct a fuzzy clustering . we suggest a new method for forecasting based on re-established a fuzzy group relation and we apply the suggest method and comparing the results of pervious methods for first and second order fuzzy relation with the number of clusters between 5 and 11 and we obtained best forecasting with m.s.e. less than the previous method.