A Comprehensive High Order Type 2 Fuzzy Time Series Forecasting Model

In this paper, we propose a hybrid high order Type 2 fuzzy time series model by combining support vector machine (SVM) with adaptive expectation model. We use SVM model to forecast the index of the fuzzy set of the predicted time. Particle swarm optimization (PSO) algorithm is used to adjust the lengths of intervals of the universe of discourse which are employed in forecasting. Moreover, we also propose a new method to calculate the weights of fuzzy sets for compensating the presence of bias in the forecasting. Further, we apply an modified adaptive model to adjust the forecasting value in the defuzzification stage. We utilize the proposed model to forecast the daily Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX) and Index 100 for the stocks and bonds exchange market of Istanbul (IMKB). The experimental results illustrate the validity of the method.