Supplementary Files
Keywords
artificial intelligence
logic programming
time-series analysis
How to Cite
Abstract
Recently, Ghiassi, Saidane and Zimbra [Int J Forecasting, vol. 21, 2005, pp. 341-362] presented a dynamicarchitecture neural network for time series prediction which performs significantly better than traditional artificial neural networks and the ARIMA methodology. The main objective of this article is to prove that the original DAN2 model can be rewritten as an additive model. We show that our formulation has several advantages: First, it reduces the total number of parameters to estimate; second, it allows estimating all the linear parameters by using ordinary least squares or ridge regression; and, finally, it improves the search for the global minimum of the error function used to estimate the model parameters. To assess the effectiveness of our approach, we estimate two models for one of the time series used as a benchmark when the original DAN2 model was proposed. The results indicate that our approach is able to find models with similar or better accuracy than the original DAN2.
GHIASSI, M. and NANGOY, S. A dynamic artificial neural network model for forecasting nonlinear processes. Computers & Industrial Engineering. 2009, vol. 57, núm. 1, pp. 287-297.
GHIASSI, M. and SAIDANE, H. A dynamic architecture for artificial neural networks. Neurocomputing. 2005, vol. 63, pp. 397-413.
GHIASSI, M.; SAIDANE, H. and ZIMBRA, D. K. A dynamic artificial neural network model for forecasting time series events. International Journal of Forecasting. 2005, vol. 21, núm. 2, pp. 341-362.
GHIASSI, M.; ZIMBRA, D.K. and SAIDANE, H. Medium term system load forecasting with dynamic artificial neural network model. Electric Power Systems Research. 2006, vol. 76, pp. 302-316.
GHIASSI, M.; ZIMBRA, D.K. and SAIDANE, H. Urban water demand forecasting with a dynamic artificial neural network model. Journal of Water Resources Planning and Management. 2008, vol. 134, núm. 2, pp. 138-146.
GOMES, G. S. S.; MAIA, A. L. S.; LUDERMIR, T. B.; CARVALHO, F. and ARAUJO, A. F. R. Hybrid model with dynamic architecture for forecasting time series. Proceedings of the International Joint Conference on Neural Networks. IEEE Computer Society. 2006, pp. 7133-7138.
GURESEN, E.; KAYAKUTLU, G. and DAIM, T.U. Using artificial neural network models in stock market index prediction. Expert Systems with Applications. 2011, vol. 38, pp. 10389-10397.
WANG, J.; NIU, D. and LI, L. Middle-long term load forecasting based on dynamic architecture for artificial neural network. Journal of Information and Computational Science. 2010, vol. 7, núm. 8, pp. 1711-1717.
VELÁSQUEZ, J. D. and FRANCO, C .J. Prediction of the prices of electricity contracts using a neuronal network with dynamic architecture. Innovar. 2010, vol. 20, núm. 36, pp. 7-14.
This journal is registered under a Creative Commons Attribution 4.0 International Public License. Thus, this work may be reproduced, distributed, and publicly shared in digital format, as long as the names of the authors and Pontificia Universidad Javeriana are acknowledged. Others are allowed to quote, adapt, transform, auto-archive, republish, and create based on this material, for any purpose (even commercial ones), provided the authorship is duly acknowledged, a link to the original work is provided, and it is specified if changes have been made. Pontificia Universidad Javeriana does not hold the rights of published works and the authors are solely responsible for the contents of their works; they keep the moral, intellectual, privacy, and publicity rights.
Approving the intervention of the work (review, copy-editing, translation, layout) and the following outreach, are granted through an use license and not through an assignment of rights. This means the journal and Pontificia Universidad Javeriana cannot be held responsible for any ethical malpractice by the authors. As a consequence of the protection granted by the use license, the journal is not required to publish recantations or modify information already published, unless the errata stems from the editorial management process. Publishing contents in this journal does not generate royalties for contributors.