Long term forecasting by combining Kohonen algorithm and standard prevision
To forecast a complete curve is a delicate problem, since the existing methods (vectorial prevision, long-term forecasting) are difficult to use and often give disappointing results. We propose a new strategy that consists in dividing up the problem into three sub-problems: prediction of the mean value and of the standard deviation and estimation of the normalized curve (the profile). The mean value and the standard deviation are predicted by any classical method (linear or neural). As to the profile, it is estimated with the help of a previous classification. The results are very convincing and a real-world application is presented : the polish electrical consumption.
Unable to display preview. Download preview PDF.
- 1.Cottrell M., Fort J.C.: esEtude d'un algorithme d'auto-organisation. Ann. Inst. Henri Poincaré, 23, 1, (1987), 1–23.Google Scholar
- 2.Cottrell M., Fort J.C. and Pagès G.: Two or three things that we know about the Kohonen algorithm. Proc. of ESANN 94, M.Verleysen Ed., Editions Quorum, (ISBN 2-9600049-1-4),(1994), 235–244.Google Scholar
- 3.Cottrell M., Girard B., Girard Y., Muller C.and Rousset P.: Daily Electrical Power Curves: Classification and Forecasting Using a Kohonen Map. From Natural to Artificial Neural Computation, Proc. IWANN'95, J.Mira, F.Sandoval eds., Lecture Notes in Computer Science, Vol.930, Springer, (1995), 1107–1113.Google Scholar
- 4.Cottrell M., Girard B., Girard Y., Mangeas M., Muller C.: Neural modeling for time series: a statistical stepwise method for weight elimination. IEEE Transactions on Neural Networks, 6, 6, (1995), 1355–1364.Google Scholar
- 5.Cottrell M., Girard B., Rousset P.: Forecasting of curves using a Kohonen classification. To appear in Journal of Forecasting, (1997).Google Scholar
- 6.Garcia Tejedor A., Cosculluela M., Bermejo C. and Montes R.: A neural system for short-term load forecasting based on day-type classification. Proc. ISAP 94, (1994).Google Scholar
- 7.Kohonen T.: Self-organized formation of topologically correct feature maps. Biol. Cyb., 43, (1982), 59–69.Google Scholar
- 8.Kohonen T.: Self-organization and associative memory. Springer, New York Berlin Heideberg, Third edition, 1989.Google Scholar
- 9.Kohonen T.: Self-Organizing Maps. Vol. 30, Springer, New York Berlin Heiderberg, 1995.Google Scholar
- 10.Lapedes A.S. and Farber R.: Non linear signal processing using neural networks: Prediction and System Modeling. Los Alamos National Laboratory Technical Report, 1987.Google Scholar
- 11.Muffler M., Cottrell M., Girard B., Girard Y., Mangeas M.: A neural network tool for forecasting French electricity consumption. Proc. WCNN 94, San Diego, (1994), I-360-365.Google Scholar
- 12.Weigend A.S., Huberman B.A. and Rumelhart D.E.: Predicting the Future: a Connectionist Approach. Int. J. Neural Systems, 1, 3, (1990), 193–209.Google Scholar
- 13.Weigend A.S. and Gershenfeld N.A.: Times Series Prediction. Proceedings Volume XV, Santa Fe Institute, Addison-Wesley, New-York, 1994.Google Scholar