Abstract
Kernel smoothing is a widely used non-parametric pattern recognition technique. By nature, it suffers from the curse of dimensionality and is usually difficult to apply to high input dimensions. In this contribution, we propose an algorithm that adapts the input metric used in multivariate regression by minimising a cross-validation estimate of the generalisation error. This allows to automatically adjust the importance of different dimensions. The improvement in terms of modelling performance is illustrated on a variable selection task where the adaptive metric kernel clearly outperforms the standard approach. Finally, we benchmark the method using the DELVE environment.
Similar content being viewed by others
References
W. Härdle, Applied Nonparametric Regression, Cambridge University Press, 1990. Econometric Society Monographs, vol. 19.
B.D. Ripley, Pattern Recognition and Neural Networks, Cambridge University Press, 1996.
H. Schiøler and U. Hartmann, “Mapping Neural Networks Derived from the Parzen Window Estimator,” Neural Networks, vol. 5, no.6, 1992, pp. 903–909.
D.F. Specht, “A General Regression Neural Network,” IEEE Transactions on Neural Networks, vol. 2, no.6, 1991, pp. 568–576.
R.E. Bellman, Adaptive Control Processes: A Guided Tour, New Jersey: Princeton University Press, 1961.
E.A. Nadaraya, “On Estimating Regression,” Theory Probab. Appl., vol. 10, 1964, pp. 186–190.
G.S. Watson, “Smooth Regression Analysis,” Sankhyā Ser. A, vol. 26, 1964, pp. 101–116.
M.P. Wand and M.C. Jones, Kernel Smoothing, London: Chapman & Hall, 1995. Monographs on Statistics and Applied Probability, vol. 60.
V.N. Vapnik, The Nature of Statistical Learning Theory, Springer, 1995.
C. Goutte, “Statistical Learning and Regularisation in Regression,” Ph.D. Thesis 1997/033, Université Paris 6, Paris, July 1997. http://eivind.imm.dtu.dk/staff/goutte/PUBLIS/thesis.html.
M. Stone, “Cross-Validatory Choice and Assessment of Statistical Predictions,” Journal of the Royal Statistical Society B, vol. 36, 1974, pp. 111–147, with discussion.
C. Goutte and J. Larsen, “Adaptive Regularization of Neural Networks Using Conjugate gradient,” in Proceedings of ICASSP'98, 1998, vol. 2, IEEE, pp. 1205–1208. ftp://eivind.imm.dtu.dk/dist/1998/goutte.icassp98.ps.Z.
R. Fletcher, Practical Methods of Optimization, Wiley, 1987.
J.H. Friedman, “Multivariate Adaptive Regression Splines,” The Annals of Statistics, vol. 19, no.1, 1991, pp. 1–141.
C. Goutte and J. Larsen, “Optimal Cross-Validation Split Ratio: Experimental Investigation,” in ICANN'98, Berlin, 1998, L. Niklasson, M. Bodén, and T. Ziemke (Eds.), vol. 2, Springer Verlag, pp. 681–686. ftp://eivind.imm.dtu.dk/dist/1998/goutte.icann98.html.
C.E. Rasmussen, R.M. Neal, G.E. Hinton, D. van Camp, M. Revow, Z. Ghahramani, R. Kustra, and R. Tibshirani, The DELVE Manual, version 1.1 edition, Dec. 1996. http://www.cs.utoronto.ca/~delve/.
C.E. Rasmussen, “Evaluation of Gaussian Processes and Other Methods for Non-Linear Regression,” PhD Thesis, Department of Computer Science, University of Toronto, 1996. http://bayes.imm.dtu.dk/.
D. Harrison and D.L. Rubinfeld, “Hedonic Prices and the Demand for Clean Air,” J. Environ. Economics and Management, vol. 5, 1978, pp. 81–102.
R.M. Neal, Bayesian Learning for Neural Networks, New York: Springer, 1996. Lecture Notes in Statistics, vol. 118.
D. Lowe, “Similarity Metric Learning for a Variable-Kernel Classifier,” Neural Computation, vol. 7, no.1, 1995, pp. 72–85.
C. Goutte and L.K. Hansen, “Regularization with a Pruning Prior,” Neural Networks, vol. 10, no.6, 1997, pp. 1053–1059.
C.K.I. Williams, “Prediction with Gaussian Processes: From Linear Regression to Linear Prediction and Beyond,” Technical Report NCRG/97/012, Neural Computing Research Group, Aston University, UK, 1997. http://neural-sever.aston.ac.uk/Papers/postscript/NCRG 97 012.ps.Z.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Goutte, C., Larsen, J. Adaptive Metric Kernel Regression. The Journal of VLSI Signal Processing-Systems for Signal, Image, and Video Technology 26, 155–167 (2000). https://doi.org/10.1023/A:1008159803952
Published:
Issue Date:
DOI: https://doi.org/10.1023/A:1008159803952