NEURON OPTIMIZATION OF EVOLUTIONARY ARTIFICIAL NEURAL NETWORKS FOR STOCK PRICE INDEX PREDICTION

Abstract

This study presents an optimization procedure for the number of processing elements (neurons) of hidden layers to predict a stock price index using Evolutionary Artificial Neural Networks (EANN), in particular, for the Istanbul Stock Market price index (ISE) in order to contribute to the development of Intelligent Systems Methods for modeling several systems that are highly non-linear and uncertain. The US dollars/Turkish Lira (US/TRY) exchange rate, Euro/Turkish Lira (EUR/TRY) exchange rate, ISE National 100 (XU100) index, world oil price, and gold price were used as for a period of approximately 10 years’ daily data as inputs. Performance is benchmarked by mean squared error, normalized mean squared error; mean absolute error and the correlation coefficient. With the fixed neural network architecture and optimized parameters, evolutionary neural networks perform better performance values when the number of neurons used in hidden layers is optimized.

Keywords

• Evolutionary algorithms, artificial neural networks, prediction, stock price index

References

1. Alkaya A. & Bayhan, G.M. (2009). The Classification of a Simulation Data of a Servo System via Evolutionary Artificial Neural Networks, International Conference on Intelligent Computing Proceedings, pp 48-54.
2. Ang J.H., Tan K.C. & Al-Mamun A. (2008). Training neural networks for classiŞcation using growth probability-based evolution, Neurocomputing ,71 3493–3508
3. Blum, A., (1992). Neural networks in C++: an object-oriented framework for building connectionist systems, John Wiley & Sons, Inc., pp. 86-103
4. Castillo-Valdivieso, P. A., Merelo J. J., & Prieto A. (2002). Statistical Analysis of the Parameters of a Neuro-Genetic Algorithm, IEEE Transactıons On Neural Networks, Vol. 13, No. 6.
5. Fine,T.L. (1999). Feedforward Neural Network Methodology, Springer, New York, pp. 129-194
6. Freitas A. (2002). A Survey of Evolutionary Algorithms for Data Mining and Knowledge, Advances in Evolutionary Computation, 2002 – Citeseer
7. Goldberg, D.E.(1989). Genetic algorithms in search, optimization and machine learning, Addison-Wesley
8. Mehrotra, K., Mohan, C. K. & Ranka, S. (1997). Elements of ArtiŞcial Neural Networks, MIT Press, Cambridge, MA.

9. Siebel, N. T., Krause, J., & Sommer, G. (2007). Efficient Learning of Neural Networks with Evolutionary Algorithms, Lecture Notes in Computer Science , Springer
10. Stepniewski, S.W. & Keane, A. J. (2006). Topology design of feedforward neural networks by genetic algorithms ,Lecture Notes in Computer Science, Springer
11. Wang, C. And Principe, J. C. (1999). Training Neural Networks With Additive Noise in The Desired Signal, IEEE Transactions on Neural Networks.
12. http://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=RBRTE&f=D
13. http://www.gold.org/investment/statistics/gold_price_chart