On multivariate fuzzy time series analysis and forecasting

Abstract

In this paper, we propose an integrated procedure for multivariate fuzzy time series modeling and its theory structure through fuzzy relation equations in this research. Combining the data of closing price and trading volume, we apply this method to construct multivariate fuzzy time series model for Taiwan Weighted Stock Index and forecast future trend while comparing the forecasting performance by average forecasting accuracy. We strongly believe that this model will be profound of meaning in forecasting future trend of financial market.