Distance rejection in the context of electric power system security assessment based on automatic learning

  • Isabelle Houben
  • Louis Wehenkel
Rejection in Pettern Recognition
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1451)

Abstract

Automatic learning methods have proved to be quite attractive in the context of power system security assessment. The complementarity of various methods proposed so far, lead us to combine them in a toolbox in order to exploit their advantages and discard their limitations. In this paper, we show how the nearest neighbor approach could be used to face the problem of detecting outliers, i.e. cases not well enough represented in the data base used to learn the models. More precisely, such detection can be based on distance rejection which implies the choice of an appropriate distance. On a particular real life problem, we show how the simple nearest neighbor in the candidate attributes space allows to reject such cases.

Keywords

power system security assessment automatic learning nearest neighbor distance rejection 

References

  1. 1.
    L. Wehenkel. Automatic learning techniques in power systems. Kluwer Academic, Boston, 1998.Google Scholar
  2. 2.
    J. R. Quinlan. Induction of decision trees. Machine Learning 1, pp. 81–106, 1986.Google Scholar
  3. 3.
    L. Breiman, J. H. Friedman, R. A. Olshen, and C. J. Stone. Classification and Regression Trees. Wadsworth International (California), 1984.Google Scholar
  4. 4.
    X. Boyen and L. Wehenkel. Automatic induction of fuzzy decision trees and its application to power system security assessment. To appear in Fuzzy Sets and Systems, 1998.Google Scholar
  5. 5.
    S. Haykin. Neural networks. A comprehensive foundation. IEEE Press, 1994.Google Scholar
  6. 6.
    R. O. Duda and P. E. Hart. Pattern classification and scene analysis. John Wiley and Sons, 1973.Google Scholar
  7. 7.
    P. A. Devijver and J. Kittler. Pattern Recognition: A Statistical Approach. Prentice-Hall International, 1982.Google Scholar
  8. 8.
    I. Houben, L. Wehenkel, and M. Pavella. Genetic algorithm based k nearest neighbors. In Proc. CIS-97, IFAC Conf. on Contr. of Indust. Syst., pp. 43–48, Belfort, Fr, 1997.Google Scholar
  9. 9.
    I. Houben, L. Wehenkel, and M. Pavella. Hybrid adaptive nearest neighbor approaches to dynamic security assessment. In Proc. of CPSPP'97, pp. 685–690, Beijing, 1997. International Academic Publishers.Google Scholar
  10. 10.
    L. Wehenkel, 1. Houben, M. Pavella, L. Riverin, and G. Versailles. Automatic learning approaches for on-line transient stability preventive control of the Hydro-Québec system — Part 1. Decision tree approaches. In Proc. of SIPOWER'95, 2nd IFAC Symp. on Control of Power Plants and Power Syst., pp. 231–236, Dec. 1995.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1998

Authors and Affiliations

  • Isabelle Houben
    • 1
  • Louis Wehenkel
    • 2
  1. 1.University of Liége - Sart-TilmanLiegeBelgium
  2. 2.Research Associate, F.N.R.S.France

Personalised recommendations