Design and Applications of Daily Trading Report Based on Gene Expression Programming with Decision Tree
| Autor: | LIN, TA-WEI, 林大為 |
|---|---|
| Rok vydání: | 2016 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 104 The purpose of this paper is to develop how GEP-DT, Gene Expression Programming with Decision Tree, apply in daily trading report at Taiwan stock market. How to find out the best variable and the optimal value of variable by Random Numerical Constants (RNC) in GEP-DT model. In addition, daily trading report are divided into five groups in different capital because this research want to know the relationship between stock price and capital. Results show that: (1) GEP-DT with RNC can find out the best variable and the optimal value of variable. At the same time, GEP-DT with RNC deal with many and complex decision variables and parameters. (2) The greater the share capital, the better the fitness value of the best. On the contrary, The smaller the share capital, the worse the value of its best fitness value. (3)After Compare internal node of GEP-DT and three kinds of select attributes methods, CfsSubset, ChiSquared and Information Gain, net branches of buy and sell is key variables. (4) GEP-DT with RNC have compared more efficient and reliable with manual processing of data or experience. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|