Within the framework of a time-series prediction competition has been held. The data to be predicted were available for the competition from Nov. 1997 till april 1998. 17 entries have been submitted before the deadline. Participants in the competition were asked to submit their predicted data together with a short description and references of the methods used. In order to stimulate wide participation in the competition, attendance of the workshop was not mandatory but was of course encouraged.

The data consists of 2000 points of ASCII data, which one could find on ftp://ftp.esat.kuleuven.ac.be/pub/sista/suykens/workshop/datacomp.dat and is plotted below.

The task was to predict the next 200 points in the time-series, and send them (also in ASCII format) to Johan.Suykens@esat.kuleuven.ac.be. The submitted data were evaluated and compared with respect to the mean squared error of the 200 predicted data points.

In order to find some good inspiration, interested participants were encouraged to consult the book `Time series prediction: forecasting the future and understanding the past' (Eds. A. Weigend & N. Gershenfeld - Addison Wesley, 1994) about a competition which was held at the Santa Fe institute.

The results of the competition and the origin of the data are described in the book

J.A.K. Suykens and J. Vandewalle (Eds.), Nonlinear Modeling: advanced black-box techniques, Kluwer Academic Publishers Boston, June 1998 (ISBN 0-7923-8195-5).The winner of the competition is James McNames (Stanford University, USA). His method is described in

J. McNames, A nearest trajectory strategy for time series prediction, Proceedings of the International Workshop on Advanced Black-Box Techniques for Nonlinear Modeling, July 8-10, 1998, K.U. Leuven Belgium, pp.112-128.and

J. McNames, J. Suykens, J. Vandewalle, Winning Entry of the K.U. Leuven time-series prediction competition, International Journal of Bifurcation and Chaos, Vol.9, No.8, pp.1485-1500, August 1999.See also

Suykens J.A.K., Vandewalle J., The K.U.Leuven competition data : a challenge for advanced neural network techniques, in Proc. of the European Symposium on Artificial Neural Networks (ESANN'2000), Bruges, Belgium, 2000, pp. 299-304.The 200 data points which had to be predicted together with a further continuation can be transfered from ftp://ftp.esat.kuleuven.ac.be/pub/sista/suykens/workshop/rescontdatacomp.dat.

This page is maintained by Johan Suykens and is last modified on August 3, 2001.