Technical Indication Generation = Trend Classification + Genetic Algorithm
| dc.contributor.author | YAP, Chur Jiun, Bernard | en_US |
| dc.contributor.author | KHOO, Siau-Cheng | en_US |
| dc.date.accessioned | 2005-01-26T03:26:12Z | en_US |
| dc.date.accessioned | 2017-01-23T06:59:41Z | |
| dc.date.available | 2005-01-26T03:26:12Z | en_US |
| dc.date.available | 2017-01-23T06:59:41Z | |
| dc.date.issued | 2005-01-26T03:26:12Z | en_US |
| dc.description.abstract | Technical indicators are used to interpret stock market trends and make investment decisions. The main diffculty of its usage lies in the fact that it is extremely hard to reliably construct an instance of any technical indicators with appropriate parameters for a particular series of stock prices. While genetic algorithm has been used to discover technical indicator in the past, not much justification have been given for such an approach. In this work, we look into the basic assumptions of the working of technical indicators, and the suitability of applying genetic algorithms for indicator discovery. We argue for the use of technical indicator as an approximation to the ideal solutions for daily trading. Consequently, we propose a framework that applies genetic algorithm to discover good parameters for any technical indicators to work well on their respective targeted price trends. A widely used technical indicator, Exponential Moving Average Crossover, is used to illustrate how the proposed framework may be employed to construct good instances of the technical indicator for a targeted price trend. | en_US |
| dc.format.extent | 318387 bytes | en_US |
| dc.format.mimetype | application/pdf | en_US |
| dc.identifier.uri | https://dl.comp.nus.edu.sg/xmlui/handle/1900.100/1523 | en_US |
| dc.language.iso | en | en_US |
| dc.relation.ispartofseries | TRC1/05 | en_US |
| dc.title | Technical Indication Generation = Trend Classification + Genetic Algorithm | en_US |
| dc.type | Technical Report | en_US |