Abstract
In this paper, a hybrid model of multy-layer neural network combined with a finite-impulse-response filter is proposed for a nonlinear time series prediction. We introduce an important analysis of the input sequence to determine the effective minimum combination of the input samples and hidden neurons. Through computer simulations, using both sunspot and computer generated time series, the proposed analysis has shown its effectiveness and the proposed predictor has demonstrated its superiority. It is of a faster convergence and smaller residual error than the conventional nonlinear predictor.
Preview
Unable to display preview. Download preview PDF.
References
A. S. Weigend and D. E. Rumrlhart: Generalization through Minimal Networks with Application to Forecasting. Proceeding of INTERFACE'91: Computing Science and Statistics, edited by Elaine Keramindas. Springer Verlag (1992) 362–370
Kinouchi, M. and Hagiwara, M.: Learning Temporal Sequences by Complex Neurons with Local Feedback. ICNN 6 (1995) 3165–3169
Haykin and L. Li: Nonlinear Adaptive Prediction of Nonstationary Signals. IEEE Trans. Signal Processing, 43 (1995) 526–535
A. S. Weigend and N. A. Gershenfeld, Time Series Prediction: Forecasting the Future and Understanding the Past. Proceedings V. XV, Santa Fe Institute (1994)
Tong, Howell and Lim, K. S.: Threshold Autoregression, Limit Cycles and Cyclical Data. Journal Royal Statistical Society B 42, (1980) 245–292
Ashraf A.M.Khalaf, and K.Nakayama: A Hybrid Neural Predictor and Its Convergence Analysis. Proceeding of the 10th Karuizawa Workshop on Circuits and Systems, Japan (1997) 357–362
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Khalaf, A.A., Nakayama, K., Hara, K. (1997). A Neural—FIR predictor: Minimum size estimation based on nonlinearity analysis of input sequence. In: Gerstner, W., Germond, A., Hasler, M., Nicoud, JD. (eds) Artificial Neural Networks — ICANN'97. ICANN 1997. Lecture Notes in Computer Science, vol 1327. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0020291
Download citation
DOI: https://doi.org/10.1007/BFb0020291
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-63631-1
Online ISBN: 978-3-540-69620-9
eBook Packages: Springer Book Archive