An Identity for Kernel Ridge Regression

  • Fedor Zhdanov
  • Yuri Kalnishkan
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6331)

Abstract

This paper provides a probabilistic derivation of an identity connecting the square loss of ridge regression in on-line mode with the loss of a retrospectively best regressor. Some corollaries of the identity providing upper bounds for the cumulative loss of on-line ridge regression are also discussed.

Keywords

Ridge Regression Kernel Matrix Reproduce Kernel Hilbert Space Prediction Output Gaussian Process Model 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Aronszajn, N.: La théorie des noyaux reproduisants et ses applications. Première partie. Proceedings of the Cambridge Philosophical Society 39, 133–153 (1943)CrossRefMathSciNetGoogle Scholar
  2. 2.
    Azoury, K.S., Warmuth, M.K.: Relative loss bounds for on-line density estimation with the exponential family of distributions. Machine Learning 43, 211–246 (2001)MATHCrossRefGoogle Scholar
  3. 3.
    Beckenbach, E.F., Bellman, R.E.: Inequalities. Springer, Heidelberg (1961)Google Scholar
  4. 4.
    Busuttil, S., Kalnishkan, Y.: Online regression competitive with changing predictors. In: Proceedings of Algorithmic Learning Theory, 18th International Conference, pp. 181–195 (2007)Google Scholar
  5. 5.
    Cesa-Bianchi, N., Long, P., Warmuth, M.K.: Worst-case quadratic loss bounds for on-line prediction of linear functions by gradient descent. IEEE Transactions on Neural Networks 7, 604–619 (1996)CrossRefGoogle Scholar
  6. 6.
    Cesa-Bianchi, N., Lugosi, G.: Prediction, Learning, and Games. Cambridge University Press, Cambridge (2006)MATHCrossRefGoogle Scholar
  7. 7.
    Henderson, H.V., Searle, S.R.: On deriving the inverse of a sum of matrices. SIAM Review 23(1) (1981)Google Scholar
  8. 8.
    Herbster, M., Warmuth, M.K.: Tracking the best linear predictor. Journal of Machine Learning Research 1, 281–309 (2001)MATHCrossRefMathSciNetGoogle Scholar
  9. 9.
    Hoerl, A.E.: Application of ridge analysis to regression problems. Chemical Engineering Progress 58, 54–59 (1962)Google Scholar
  10. 10.
    Kakade, S.M., Seeger, M.W., Foster, D.P.: Worst-case bounds for Gaussian process models. In: Proceedings of the 19th Annual Conference on Neural Information Processing Systems (2005)Google Scholar
  11. 11.
    Kivinen, J., Warmuth, M.K.: Exponentiated gradient versus gradient descent for linear predictors. Infornation and Computation 132(1), 1–63 (1997)MATHCrossRefMathSciNetGoogle Scholar
  12. 12.
    Kumon, M., Takemura, A., Takeuchi, K.: Sequential optimizing strategy in multi-dimensional bounded forecasting games. CoRR abs/0911.3933v1 (2009)Google Scholar
  13. 13.
    Lamperti, J.: Stochastic Processes: A Survey of the Mathematical Theory. Springer, Heidelberg (1977)MATHGoogle Scholar
  14. 14.
    Rasmussen, C.E., Williams, C.K.I.: Gaussian Processes for Machine Learning. MIT Press, Cambridge (2006)MATHGoogle Scholar
  15. 15.
    Saunders, C., Gammerman, A., Vovk, V.: Ridge regression learning algorithm in dual variables. In: Proceedings of the 15th International Conference on Machine Learning, pp. 515–521 (1998)Google Scholar
  16. 16.
    Seeger, M.W., Kakade, S.M., Foster, D.P.: Information consistency of nonparametric Gaussian process methods. IEEE Transactions on Information Theory 54(5), 2376–2382 (2008)CrossRefMathSciNetGoogle Scholar
  17. 17.
    Vovk, V.: Competitive on-line statistics. International Statistical Review 69(2), 213–248 (2001)MATHCrossRefGoogle Scholar
  18. 18.
    Zhdanov, F., Vovk, V.: Competing with gaussian linear experts. CoRR abs/0910.4683 (2009)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2010

Authors and Affiliations

  • Fedor Zhdanov
    • 1
  • Yuri Kalnishkan
    • 1
  1. 1.Computer Learning Research Centre and Department of Computer ScienceRoyal Holloway, University of LondonEgham, SurreyUnited Kingdom

Personalised recommendations