|Title:||Prediction of a financial time series by neural networks : a simple chaotic approach|
|Subject:||Stock price forecasting -- Mathematical models|
Neural networks (Computer science)
Time-series analysis -- Mathematical models
Chaotic behavior in systems
Hong Kong Polytechnic University -- Dissertations
Department of Computing
|Pages:||xi, 95,  leaves : ill. ; 30 cm|
|Abstract:||During the 1950s and early 1960s, the Efficient Market Hypothesis emerged in which the stock prices reflect all relevant information in the markets. The independently distributed return indicates that the markets have no memory and no underlying dynamics. Some phenomena such as run up in stock prices over time and the high volatility observed in stock markets cannot be explained. In the last decade, the ability of chaotic models of generating time path which exhibits a volatile behaviour similar to that observed in financial time series becomes the focus of attention. A number of researches also claimed that there was evidence of chaotic behaviour in certain financial time series. In case a nonlinear dynamics underlies a financial index, the deterministic nonlinear model reveals that accurate short-term prediction of the index is possible. In this study, one of the most widely used tests for chaos: correlation dimension test is applied on financial indexes and currency exchange rate to study the possibility of existence of chaos. The Hong Kong Hang Seng Index, which is potential for existence of chaos, is selected for prediction. The forecasting is captured by excellent nonlinear prediction models, neural networks complementing with genetic algorithms and is proceeded in different aspects: 1. The Hang Seng Index is predicted by phase space reconstruction of the attractor. Excellent short-term prediction is expected if it is chaotic. 2. Including the time series itself, different derived indicators are predicted by recognition of the recurring patterns and nonlinear relationships within the time series. The prediction results are evaluated by different trading strategies and demonstrates the superiority of neural network technology in forecasting. The project concludes that the Hang Seng Index is not chaotic as the short-term trading result by phase space reconstruction is much inferior to that of a long-term trading model.|
|Rights:||All rights reserved|
Files in This Item:
|b14369345.pdf||For All Users (off-campus access for PolyU Staff & Students only)||4.05 MB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: