Using Support Vector Machine to Analysis the Taiwan Weighted Stock Index
| Autor: | Chun-kai Tseng, 曾俊凱 |
|---|---|
| Rok vydání: | 2015 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 103 This research applies Support Vector Machine for efficient prediction of Taiwan Weighted Stock Price Index (TAIEX). Different indicators from the technical analysis field of study are used as input features. The research method is to change the data type for the correct rate and to discuss how to select the data. To avoid the accuracy of major events affecting the classifier, it is important to decide the training period of time through the experiment, which include training period of all information or training period that events were disappeared. Considering the importance of input features, the correlation can be found between the two input features from the experiment. And by deleting unimportant variables, the classification ability of Support Vector Machines can be imporved. Finally, the distribution from the Permutation Test in comparison with the original result of classification shows original result of classification does not exist in the critical regions. This indicates excellent classification result. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|