Abstract
Motivated by the slow learning properties of multilayer perceptrons (MLPs) which utilize computationally intensive training algorithms, such as the backpropagation learning algorithm, and can get trapped in local minima, this work deals with ridge polynomial neural networks (RPNN), which maintain fast learning properties and powerful mapping capabilities of single layer high order neural networks. The RPNN is constructed from a number of increasing orders of Pi–Sigma units, which are used to capture the underlying patterns in financial time series signals and to predict future trends in the financial market. In particular, this paper systematically investigates a method of pre-processing the financial signals in order to reduce the influence of their trends. The performance of the networks is benchmarked against the performance of MLPs, functional link neural networks (FLNN), and Pi–Sigma neural networks (PSNN). Simulation results clearly demonstrate that RPNNs generate higher profit returns with fast convergence on various noisy financial signals.
Similar content being viewed by others
References
Leerink LR, Giles CL, Horne BG, Jabri MA (1995) Learning with product units. In: Tesaro G, Touretzky D, Leen T (eds) Advances in neural information processing systems 7, MIT Press, Cambridge, MA, pp 537–544
Lawrence S, Giles CL (2000) Overfitting and neural networks: conjugate gradient and back propagation. In: International joint conference on neural network, Italy, IEEE Computer Society, CA, pp 114–119
Hellström T, Holmström K (1998) Predicting the stock market. Technical Report IMa-TOM-1997-07, Center of Mathematical Modeling, Department of Mathematics and Physis, Mälardalen University, Västeras, Sweden
Dunis CL, Wiliams M (2002) Modeling and trading the UER/USD exchange rate: do neural network models perform better? In: Derivatives use, trading and regulation, vol 8, No. 3, pp 211–239
Yao J, Tan CL (2000) A case study on neural networks to perform technical forecasting of forex. Neurocomputing 34:79–98
Shachmurove Y, Witkowska D (2000) Utilizing artificial neural network model to predict stock markets. CARESS Working Paper #00–11
Cass R, Radl B (1996) Adaptive process optimization using functional-link networks and evolutionary algorithm. Control Eng Pract 4(11):1579–1584
Liatsis P, Hussain AJ (1999) Nonlinear one-dimensional DPCM image prediction using polynomial neural networks. In: Proc. SPIE, applications of artificial neural networks in image processing IV, San Jose, California, 28–29 January, vol 3647, pp 58–68
Tawfik H, Liatsis P (1997) Prediction of non-linear time-series using higher-order neural networks. In: Proceeding IWSSIP’97 conference, Poznan, Poland
Mirea L, Marcu T (2002) System identification using functional-link neural networks with dynamic structure. In: 15th Triennial world congress, Barcelona, Spain
Shin Y, Ghosh J (1991) The Pi–Sigma networks: an efficient higher-order neural network for pattern classification and function approximation. In: Proceedings of international joint conference on neural networks, Seattle, Washington, July 1991, vol 1, pp 13–18
Shin Y, Ghosh J (1992) Computationally efficient invariant pattern recognition with higher order Pi–Sigma networks. The University of Texas at Austin
Kaita T, Tomita S, Yamanaka J (2002) On a higher-order neural network for distortion invariant pattern recognition. Pattern Recognit Lett 23:977–984
Pau YH, Phillips SM (1995) The functional link net and learning optimal control. Neurocomputing 9:149–164
Artyomov E, Pecht OY (2004) Modified high-order neural network for pattern recognition. Pattern Recognit Lett 26(6):843–851
Giles CL, Maxwell T (1987) Learning, invariance and generalization in high-order neural networks. In: Applied optics, vol 26, no 23. Optical Society of America, Washington D.C., pp 4972–4978
Shin Y, Ghosh J (1995) Ridge polynomial networks. IEEE Trans Neural Netwo 6(3):610–622
Voutriaridis C, Boutalis YS, Mertzios G (2003) Ridge polynomial networks in pattern recognition. EC-VIP-MC 2003. In: 4th EURASIP conference focused on video/image processing and multimedia communications, Croatia, pp 519–524
Shin Y, Ghosh J (1992) Approximation of multivariate functions using ridge polynomial networks. In: Proceedings of international joint conference on neural networks, vol 2, pp 380–385
Karnavas YL, Papadopoulos DP (2004) Excitation control of a synchronous machine using polynomial neural networks. J Electr Eng 55(7–8):169–179
Durbin R, Rumelhart DE (1989) Product units: a computationally powerful and biologically plausible extension to back-propagation networks. Neural Comput 1:133–142
Giles CL, Griffin RD, Maxwell T (1998) Encoding geometric invariances in HONN. American Institute of Physics, USA, pp 310–309
Fei G, Yu YL (1994) A modified Sigma–Pi BP network with self-feedback and its application in time series analysis. In: Proceedings of the 5th international conference, vol 2243–508F, pp 508–515
Thimm G (1995) Optimization of high order perceptron. Swiss federal Institute of Technology (EPFL)
Ghosh J, Shin Y (1992) Efficient higher-order neural networks for function approximation and classification. Int J Neural Syst 3(4):323–350
Shin Y, Ghosh J (1992) Computationally efficient invariant pattern recognition with higher order Pi–Sigma networks. The University of Texas at Austin
Schmitt M (2001) On the complexity of computing and learning with multiplicative neural networks. Neural Comput 14:241–301
Schwaerzel R (1996) Improving the prediction accuracy of financial time series by using multi-neural network systems and enhanced data preprocessing. Thesis, Master of Science, The University of Texas at San Antonio
Knowles C, Hussain A, El Deredy W, Lisboa P (2005) Higher order neural networks for the prediction of financial time series. Forecasting Financial Markets, France
Hyndman RJ (2005) Time series data library. Downloaded from: http://www-personal.buseco.monash.edu.au/∼hyndman/TSDL/. Accessed on September 2005. Original source from: McCleary and Hay, Applied time series analysis for the social sciences, 1980, Sage Publications
Thomason M (1999) The practitioner method and tools. J Comput Intell Fin 7(3):36–45
Thomason M (1999) The practitioner method and tools. J Comput Intell Fin 7(4):35–45
Cao LJ, Francis EHT (2003) Support vector machine with adaptive parameters in financial time series forecasting. IEEE Trans Neural Netw 14:6
Haykin S (1999) Neural networks. A comprehensive foundation, 2nd edn. Prentice-Hall, New Jersey
Acknowledgments
The work of R. Ghazali is supported by Universiti Tun Hussein Onn (UTHM), Malaysia.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Ghazali, R., Hussain, A.J., Liatsis, P. et al. The application of ridge polynomial neural network to multi-step ahead financial time series prediction. Neural Comput & Applic 17, 311–323 (2008). https://doi.org/10.1007/s00521-007-0132-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-007-0132-8