Predicting stock market time series using evolutionary artificial neural networks with hurst exponent input windows

Authors:
Somesh Selvaratnam;Michael Kirley
Affiliations:
Department of Computer Science and Software Engineering, The University of Melbourne, Parkville, Victoria, Australia;Department of Computer Science and Software Engineering, The University of Melbourne, Parkville, Victoria, Australia
Venue:
AI'06 Proceedings of the 19th Australian joint conference on Artificial Intelligence: advances in Artificial Intelligence
Year:
2006

Citing 6
Cited 0

Application of neural networks to an emerging financial market: forecasting and trading the Taiwan stock index

Computers and Operations Research - Special issue: Emerging economics
Modeling chaotic behavior of stock indices using intelligent paradigms

Neural, Parallel & Scientific Computations - Special issue: Advances in intelligent systems and applications
A r/s approach to trends breaks detection

KES'05 Proceedings of the 9th international conference on Knowledge-Based Intelligent Information and Engineering Systems - Volume Part I
Forecasting time series with genetic fuzzy predictor ensemble

IEEE Transactions on Fuzzy Systems
A new evolutionary system for evolving artificial neural networks

IEEE Transactions on Neural Networks
Financial volatility trading using recurrent neural networks

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

Predicting stock market time series is a challenging problem due to their random nature, non-stationarity and noise. In this study, we introduce an enhanced evolutionary artificial neural network (EANN) model to meet this challenge. Here, fractal analyses based on Hurst exponent calculations are used to characterize the time series and to identify appropriate input windows for the EANN. We investigate the efficacy of the model using closing price time series for a suite of stocks listed on the SPI index on the Australian Stock Exchange. The results show that Hurst exponent configured models out-perform basic EANN models in terms of average trading profit found using a simple trading strategy.