Abstract
A combination of singular spectrum analysis and locally linear neurofuzzy modeling technique is proposed to make accurate long-term prediction of natural phenomena. The principal components (PCs) obtained from spectral analysis have narrow band frequency spectra and definite linear or nonlinear trends and periodic patterns; hence they are predictable in large prediction horizon. The incremental learning algorithm initiates a model for each of the components as an optimal linear least squares estimation, and adds the nonlinear neurons if they help to reduce error indices over training and validation sets. Therefore, the algorithm automatically constructs the best linear or nonlinear model for each of the PCs to achieve maximum generalization, and the long-term prediction of the original time series is obtained by recombining the predicted components. The proposed method has been primarily tested in long-term prediction of some well-known nonlinear time series obtained from Mackey–Glass, Lorenz, and Ikeda map chaotic systems, and the results have been compared to the predictions made by multi-layered perceptron (MLP) and radial basis functions (RBF) networks. As a real world case study, the method has been applied to the long-term prediction of solar activity where the results have been compared to the long-term predictions of physical precursor and solar dynamo methods.
Similar content being viewed by others
References
Nelles O (2001) Nonlinear system identification. Springer, Berlin Heidelberg New York
Haykin S (1994) Neural networks: a comprehensive foundation. Macmillan, New York
Cichoki A, Chichester R (1993) Neural networks for optimization and signal processing. Wiley, New York
Bossley KM (1997) Neurofuzzy modelling approaches in system identification. PhD thesis, University of Southampton, Southampton
Park J, Sandberg IW (1993) Approximation and radial basis function networks. Neural Comput 5:305–316
Lillekjendlie B, Kugiumtzis D, Christophersen N (1994) Chaotic time series, part II: system identification and prediction. Model Identif Control 15:225–243
Leung H, Lo T, Wang S (2001) Prediction of noisy chaotic time series using an optimal radial basis function neural network. IEEE Trans Neural Netw 12(5):1163–1172
Cao L, Hong Y, Fang H, He G (1995) Predicting chaotic time series with wavelet networks. Physica D 85:225–238
Farmer JD, Sidorowich JJ (1987) Predicting chaotic time series. Phys Rev Lett 59(8):845–848
Abarbanel HDI (1996) Analysis of observed chaotic data. Springer, Berlin Heidelberg New York
Ott E (1993) Chaos in dynamical systems. Cambridge University Press, Cambridge
Ott E, Sauer T, Yorke JA (eds) (1994) Coping with chaos: analysis of chaotic data and the exploitation of chaotic systems. Wiley, New York
Fredrich K (1986) Estimating the dimension of weather and climate attractors. J Atmos Sci 43:419–432
Medio A (1992) Chaotic dynamics: theory and applications to economics. Cambridge University Press, Cambridge
Navone HD, Ceccatto HA (1995) Forecasting chaos from small data sets: a comparison of different nonlinear algorithms. J Phys A 28(12):3381–3388
Brown GM (1992) The peak of solar cycle 22: predictions in retrospect. Ann Geophys 10:453–470
Thompson R (1993) A technique for predicting the amplitude of solar cycle. Solar Phys 148:383
Joselyn JA, et al (1997) Panel achieves consensus prediction of solar cycle. EOS Trans AGU 78:211–212
Schatten KH, Scherrer PH, Svalgaard L, Wilcox JM (1978) Using dynamo theory to predict the sunspot number during solar cycle 21. Geophys Res Lett 5:411
Schatten KH, Sofia S (1987) Forecast of an exceptionally large even numbered solar cycle. Geophys Res Lett 14:632
Schatten KH, Pesnell WD (1993) An early solar dynamo prediction: cycle 23 ∼ cycle 22. Geophys Res Lett 20:2257–2278
Schatten KH, Myers DJ, Sofia S (1996) Solar activity forecast for solar cycle 23. Geophys Res Lett 23(6):605–608
Sofia S, Fox P, Schatten KH (1998) Forecast update for activity cycle 23 from a dynamo-based method. Geophys Res Lett 25(22):4149–4152
Tong H, Lim K (1980) Threshold autoregressive limit cycles and cyclical data. J R Stat Soc B 42:245–292
Weigend A, Huberman B, Rumelhart DE (1992) Predicting sunspots and exchange rates with connectionist networks. In: Casdagli E (eds) Nonlinear modeling and forecasting. Addison–Wesley, Reading, pp 395–432
Uluyol O, Ragheb M, Ray SR (1998) Local output gamma feedback neural network. Proc IEEE Int Conf Neural Netw IJCNN 1:337–342
Tong H (1996) Nonlinear time series: a dynamical system approach. Oxford University Press, London
Lucas C, Abbaspour A, Gholipour A, Araabi BN, Fatourechi M (2003) Enhancing the performance of neurofuzzy predictors by emotional learning algorithm. Informatica 27(2):165–174
Gholipour A, Abbaspour A, Araabi BN, Lucas C (2003) Enhancements in the prediction of solar activity by locally linear model tree. Proceedings of MIC2003: 22nd international conference on modeling, identification and control. Innsbruck, pp 158–161
Vautard R, Ghil M (1989) Singular spectrum analysis in nonlinear dynamics with applications to paleoclimatic time series. Physica D 35:395–424
Vautard R, Yiou P, Ghil M (1992) Singular spectrum analysis: a toolkit for short noisy chaotic signals. Physica D 58:95–126
Kugiumtzis D, Lillekjendlie B, Christophersen N (1994) Chaotic time series, part I: Estimation of some invariant properties in state space. Model Identif Control 15(4):205–224
Mackey M, Glass L (1977) Oscillation and chaos in physiological control systems. Science 197:281–287
Fröyland J (1992) Introduction to chaos and coherence. IOP Publishing Ltd, London
So P, Ott E, Sauer T, Gluckman B, Grebogi C, Schiff S (1997) Extracting unstable periodic orbits from chaotic time series data. Phys Rev E 55(5):5398–5417
Chatfield C (1989) The analysis of time series: an introduction. Chapman and Hall, London
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Gholipour, A., Lucas, C., Araabi, B.N. et al. Extracting the main patterns of natural time series for long-term neurofuzzy prediction. Neural Comput & Applic 16, 383–393 (2007). https://doi.org/10.1007/s00521-006-0062-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-006-0062-x