Fuzzy time series and its models
Fuzzy Sets and Systems
Forecasting enrollments with fuzzy time series—part I
Fuzzy Sets and Systems
Forecasting enrollments with fuzzy time series—part II
Fuzzy Sets and Systems
Forecasting enrollments based on fuzzy time series
Fuzzy Sets and Systems
Adaptive learning defuzzification techniques and applications
Fuzzy Sets and Systems
Handling forecasting problems using fuzzy time series
Fuzzy Sets and Systems
Deterministic fuzzy time series model for forecasting enrollments
Computers & Mathematics with Applications
Fuzzy relation analysis in fuzzy time series model
Computers & Mathematics with Applications
Temperature prediction using fuzzy time series
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
Handling forecasting problems based on two-factors high-order fuzzy time series
IEEE Transactions on Fuzzy Systems
Hi-index | 0.00 |
This study uses two sets of Taiwanese data, the export values as the prediction varialbe and its foreign exchange spot rates as the auxiliary variable, to discuss two important issues of forecasting effects in the fuzzy time series analysis by using One- and Two-factor models. The first issue is the relation between the optimum number of partition equal intervals and forecasting error. The second issue is the setting of fuzzy matrix (Bi) in the model to compare its impacts on forecasting error when it is static or dynamic. The above two issues are investigated with the empirical results. First, the optimum number of partition equal intervals is to select 14 intervals for the information to have the smallest forecasting error in all models for one- or twofactor, or different number of window basis selected. However, if partitioning the information into more than 14 equal intervals, the forecasting error can not be reduced but presents a waving pattern. Second, when the information period is longer and if the selecting window basis is two, under any number of partition intervals, the forecasting error is always smaller for the dynamic Bi than for the static one. However, when the information period is shorter and the window basis is two or three, only partitioning into five or eight equal intervals, the forecasting error will also be smaller for the dynamic Bi.