| Summary: | 碩士 === 國立臺灣科技大學 === 資訊工程系 === 93 === In recent years, many researchers used fuzzy time series to handle prediction problems. It is obvious that an event may be affected by many factors. For dealing with forecasting problems, if we consider more factors for prediction, then we can get better forecasting results. In this thesis, we present three new methods for dealing with forecasting problems, based on genetic algorithms, simulated annealing algorithms and high-order fuzzy time series. In the first method, we present a new method to predict the temperature and the TAIFEX (Taiwan Futures Exchange), based on the two-factors high-order fuzzy time series. The proposed method constructs two-factors high-order fuzzy logical relationships based on the historical data to increase the forecasting accuracy rate. In the second method, we present a new method for temperature prediction and the TAIFEX forecasting based on genetic algorithms and two-factors high-order fuzzy time series. The proposed method constructs two-factors high-order fuzzy logical relationship based on the historical data and uses genetic algorithms to adjust the length of each interval in the universe of discourse for temperature prediction and the TAIFEX forecasting to increase the forecasting accuracy rate. In the third method, we improve the second method to present a new method for temperature prediction and the TAIFEX forecasting based on genetic simulated annealing techniques and high-order fuzzy time series, where the simulated annealing techniques are used to deal with the mutation operations of genetic algorithms and can avoid falling into the local optimum effectively for increasing the forecasting accuracy rate. The proposed methods get higher forecasting accuracy rates than the existing methods.
|
|---|
