This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
On the Optimal Parameter Choice for v-Support Vector Machines
October 2003 (vol. 25 no. 10)
pp. 1274-1284

Abstract—We determine the asymptotically optimal choice of the parameter ν for classifiers of ν-support vector machine (ν-SVM) type which has been introduced by Schölkopf et al. It turns out that ν should be a close upper estimate of twice the optimal Bayes risk provided that the classifier uses a so-called universal kernel such as the Gaussian RBF kernel. Moreover, several experiments show that this result can be used to implement some modified cross validation procedures which improve standard cross validation for ν-SVMs.

[1] C.-C. Chang and C.-J. Lin, Training Nu-Support Vector Classifiers: Theory and Algorithms, Neural Computation, vol. 13, pp. 2119-2147, 2001.
[2] C.-C. Chang and C.-J. Lin LIBSVM 2.33,http://www.csie.ntu. du.twcjlin/. 2002.
[3] N. Cristianini, and J. Shawe-Taylor, Support Vector Machines and Other Kernel-Based Learning Methods. Cambridge Univ. Press, 2000.
[4] F. Cucker and S. Smale, On the Mathematical Foundations of Learning Bulletin Am. Math. Soc., vol. 39, pp. 1-49, 2002.
[5] W. Dahmen and C.A. Micchelli, Some Remarks on Ridge Functions Approximate Theory Application, vol. 3, pp. 139-143, 1987.
[6] L. Devroye, L. Györfi, and G. Lugosi, A Probabilistic Theory of Pattern Recognition. New York: Springer, 1997.
[7] R.M. Dudley, A Course on Empirical Processes Lecture Notes in Math., vol. 1097, pp. 1-142, 1984.
[8] R.M. Dudley, Real Analysis and Probability. New York: Chapman Hall, 1989.
[9] T. Evgeniou, M. Pontil, and T. Poggio, Regularization Networks and Support Vector Machines Advances in Computational Mathematics, vol. 13, pp. 1-50, 2000.
[10] M. Loève, Probability Theory, third ed., Princeton: D. van Nostrand, 1963.
[11] G. Rätsch http://mlg.anu.edu.au/~raetsch/dataindex.html. 2003.
[12] G. Rätsch, T. Onoda, and K.-R. Müller, Soft Margins for AdaBoost Machine Learning, vol. 42, pp. 287-320, 2001.
[13] G. Rätsch, Oral Communication.
[14] B. Schölkopf, A.J. Smola, R.C. Williamson, and P.L. Bartlett, New Support Vector Algorithms Neural Computation, vol. 12, pp. 1207-1245, 2000.
[15] B. Schölkopf and A.J. Smola, Learning with Kernels. MIT Press, 2001.
[16] I. Steinwart, On the Influence of the Kernel on the Consistency of Support Vector Machines J. Machine Learning Research, vol. 2, pp. 67-93, 2001.
[17] I. Steinwart, Support Vector Machines Are Universally Consistent J. Complexity, vol. 18, pp. 768-791, 2002.
[18] I. Steinwart, Consistency of Support Vector Machines and Other Regularized Kernel Machines to be published in IEEE Trans. Information Theory, pending publication.
[19] I. Steinwart, Sparseness of Support Vector Machines J. Machine Learning Research, to be published, 2003

Index Terms:
Pattern recognition, PAC model, support vector machines, parameter selection.
Citation:
Ingo Steinwart, "On the Optimal Parameter Choice for v-Support Vector Machines," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 10, pp. 1274-1284, Oct. 2003, doi:10.1109/TPAMI.2003.1233901
Usage of this product signifies your acceptance of the Terms of Use.