| Summary: | 碩士 === 輔仁大學 === 管理學研究所 === 94 === This study predicts the opening prices and closing prices of TAIFEX index futures by the information of international stock price index, TAIFEX index and the technical indicators of TAIFEX. The purpose of this study is to investigate if the two-stage hybrid model of artificial neural network results performs better than the other models by neural network models with various sources of information. In addition, the other purpose is to establish the transaction strategy to compare the profit by using various models, with different sources of information.
The sample period of the study is from October 2, 2004 to February 27, 2006, with 597 transaction dates, the results are as follows:
1. It is revealed that the prediction by the model with the indicators of technical analysis of TAIFEX information is better than the other two models with international stock index price information, and TAIFEX information, respectively.
2. The result reveals that the prediction by two–stage hybrid model is better than those single–stage models.
3. The transaction strategy by two-stage hybrid model makes more profit than the rest of the models.
|
|---|
