Abstract
We present a recurrent neural network topology, the Shared Layer Percep-tron, which allows robust forecasts of complex systems. This is achieved by several means. First, the forecasts are multivariate, i. e., all observables are forecasted at once. We avoid overfitting the network to a specific observable. The output at time step t, serves as input for the forecast at time step t+1. In this way, multi step forecasts are easily achieved. Second, training several networks allows us to get not only a point forecast, but a distribution of future realizations. Third, we acknowledge that the dynamic system we want to forecast is not isolated in the world. Rather, there may be a multitude of other variables not included in our analysis which may influence the dynamics. To accommodate this, the observable states are augmented by hidden states. The hidden states allow the system to develop its own internal dynamics and harden it against external shocks. Relatedly, the hidden states allow to build up a memory. Our example includes 25 financial time series, representing a market, i. e., stock indices, interest rates, currency rates, and commodities, all from different regions of the world. We use the Shared Layer Perceptron to produce forecasts up to 20 steps into the future and present three applications: transaction decision support with market timing, value at risk, and a simple trading strategy.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Michael H. Breitner, Frank Köller, Simon König, and Hans-Jörg von Mettenheim. Intelligent decision support systems and neurosimulators: A promising alliance for financial services providers. In H. Österle, J. Schelp, and R. Winter, editors, Proceedings of the European Conference on Information Systems (ECIS) 2007, St. Gallen, pages 478–489, 2007.
Michael H. Breitner, Corinna Luedtke, Hans-Jörg von Mettenheim, Daniel Rösch, Philipp Sibbertsen, and Grigoriy Tymchenko. Modeling portfolio value at risk with statistical and neural network approaches. In Christian Dunis, Michael Dempster, Michael H. Breitner, Daniel Rösch, and Hans-Jörg von Mettenheim, editors, Proceedings of the 17th International Conference on Forecasting Financial Markets: Advances for Exchange Rates, Interest Rates and Asset Management, Hannover, 26–28 May 2010, 2010.
Christian Dunis, Jason Laws, and Patrick Naïm. Applied quantitative methods for trading and investment. Wiley, Southern Gate, Chichester, 2003.
Christian L. Dunis, Jason Laws, and Georgios Sermpinis. Higher order and recurrent neural architectures for trading the EUR/USD exchange rate. Quantitative Finance, 10, 2010.
John C. Hull. Options, Futures and Other Derivatives. Prentice Hall, 6th edition, 2006.
Ashraf Laïdi. Currency Trading and Intermarket Analysis. Wiley, New Jersey, 2009.
Hans-Jörg von Mettenheim and Michael H. Breitner. Robust forecasts with shared layer per-ceptrons. In Christian Dunis, Michael Dempster, Michael H. Breitner, Daniel Rösch, and Hans-Jörg von Mettenheim, editors, Proceedings of the 17th International Conference on Forecasting Financial Markets: Advances for Exchange Rates, Interest Rates and Asset Management, Hannover, 26–28 May 2010, 2010.
Hans Georg Zimmermann. Forecasting the Dow Jones with historical consistent neural networks. In Christian Dunis, Michael Dempster, and Virginie Terraza, editors, Proceedings of the 16th International Conference on Forecasting Financial Markets: Advances for Exchange Rates, Interest Rates and Asset Management, Luxembourg, 27–29 May 2009, 2009.
Hans Georg Zimmermann. Advanced forecasting with neural networks. In Christian Dunis, Michael Dempster, Michael H. Breitner, Daniel Rösch, and Hans-Jörg von Mettenheim, editors, Proceedings of the 17th International Conference on Forecasting Financial Markets: Advances for Exchange Rates, Interest Rates and Asset Management, Hannover, 26–28 May 2010, 2010.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
von Mettenheim, HJ., Breitner, M.H. (2011). Forecasting Complex Systems with Shared Layer Perceptrons. In: Hu, B., Morasch, K., Pickl, S., Siegle, M. (eds) Operations Research Proceedings 2010. Operations Research Proceedings. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-20009-0_3
Download citation
DOI: https://doi.org/10.1007/978-3-642-20009-0_3
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-20008-3
Online ISBN: 978-3-642-20009-0
eBook Packages: Business and EconomicsBusiness and Management (R0)