| Summary: | 碩士 === 國立臺灣科技大學 === 資訊工程系 === 103 === In this thesis, we propose two new fuzzy forecasting methods to deal with forecasting problems based on fuzzy logical relationships, fuzzy-trend logical relationship groups, K-means clustering algorithm, similarity measures and particle swarm optimization (PSO) techniques. In the first method of this thesis, we propose a new method for forecasting the Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX) based on fuzzy time series, fuzzy logical relationships, particle swarm optimization (PSO) techniques, the K-means clustering algorithm, and similarity measures between the subscripts of the fuzzy sets, where we use PSO techniques to get the optimal partition of the intervals in the universe of discourse, use the K-means clustering algorithm to cluster the subscripts of the fuzzy sets of the current states of the fuzzy logical relationships to get the cluster center of each cluster and to divide the constructed fuzzy logical relationships into fuzzy logical relationship groups, and use similarity measures between the subscripts of the fuzzy sets for forecasting the TAIEX. In the second method of this thesis, we propose a new method for forecasting the TAIEX and the NTD/USD exchange rates based on two-factors second-order fuzzy-trend logical relationship groups, PSO techniques and similarity measures between the subscripts of fuzzy sets, where we use PSO techniques to get the optimal partition of the intervals in the universe of discourse, use similarity measures between the subscripts of the fuzzy sets to choose fuzzy logical relationships from the fuzzy-trend logical relationship group, and use the probabilities of trends for forecasting the TAIEX and the NTD/USD exchange rates. The experimental results show that the proposed methods get higher forecasting accuracy rates than the existing methods.
|
|---|
