| Presentation | 2004/1/19 Evolution of Investment Strategy in an Environment of Market Represented by Financial Time Series (Neurocomputing) Mieko Tanaka-Yamawaki, Tomohiro Motoyama, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | We attempt to generate investment strategy in computer simulations by regarding each strategy as an agent, that evolves to the next generation according to the genetic mechanism. Varying environment created by other agents in a competitive financial market is assumed to be represented by the financial time series measured tick by tick. We first consider predicting the direction of the next motion of tick-wise price fluctuation based on the information of the past direction of motion (Sign of increment method), then see if the prediction improves if we increase the length of history from one tick to 7 ticks. We obtained a natural result of increasing performance for longer history for the history length less than five. However the performance saturates after five ticks and the best hitting rate does not go beyond the level of 70%. We then considered incorporating different source of input, such as the level of the absolute value of the current price (not a price difference) compared to the average of the absolute values of past N ticks (Relative price method), which did not show any better performance. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Evolutional algorithm / Agent simulation / Investment strategy / Tick data / Financial time series / Sign of increment / Relative prices |
| Paper # | NC2003-110 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 2004/1/19(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Neurocomputing (NC) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Evolution of Investment Strategy in an Environment of Market Represented by Financial Time Series (Neurocomputing) |
| Sub Title (in English) | |
| Keyword(1) | Evolutional algorithm |
| Keyword(2) | Agent simulation |
| Keyword(3) | Investment strategy |
| Keyword(4) | Tick data |
| Keyword(5) | Financial time series |
| Keyword(6) | Sign of increment |
| Keyword(7) | Relative prices |
| 1st Author's Name | Mieko Tanaka-Yamawaki |
| 1st Author's Affiliation | Faculty of Engineering, Tottori University() |
| 2nd Author's Name | Tomohiro Motoyama |
| 2nd Author's Affiliation | Faculty of Engineering, Tottori University |
| Date | 2004/1/19 |
| Paper # | NC2003-110 |
| Volume (vol) | vol.103 |
| Number (no) | 601 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |