The Seasonal Adjustment Of Seasonal Time Series And Fuzzy Forecasting Method

Fuzzy time series models have been developed during the last decade. These models usually do not consider time series stationarity in prediction. However, if the time series is not stationarity, it will affect the prediction accuracy. Meanwhile, several factors that affect the forecasting accuracy, such as universe of discourse partition and data fuzzification are two particularly important parties on the fuzzy time series models, the above problems will be studied in this paper to develop an improved fuzzy time series model for forecasting the seasonal pattern of data.In the seasonal adjustment. In view of the problems that using the ratio of the moving average algorithm to seasonal adjustment will miss some items. Based on this, the paper improves this method to analysis trend, the time series will become either stationary or trend.In the universe of discourse partition, the paper employ the FCM algorithm to partition the universe of discourse and proposes one kind of validity standard to judge the clustering effect and improve the accuracy of prediction.In data fuzzification, a subjective definition method cannot really reflect the distribution of data structure. The paper use the chen’s definition of fuzzy sets base on the distance.Finally by the prediction results of Taiwan machine production data shows that the use of this new fuzzy time series model is effective.