论文信息 - Effects of norms on learning properties of support vector machines

Effects of norms on learning properties of support vector machines

Support vector machines (SVMs) are known to have a high generalization ability, yet a heavy computational load since margin maximization results in a quadratic programming problem. It is known that this maximization task results in a pth-order programming problem if we employ the L/sub P/ norm instead of the L/sub 2/ norm. In this paper, we theoretically show the effects of p on the learning properties of SVMs by clarifying its geometrical meaning.

Kazushi Ikeda | Noboru Murata

[1] Noboru Murata,et al. Support vector machines with different norms: motivation, formulations and results , 2001, Pattern Recognit. Lett..

[2] Olvi L. Mangasarian,et al. Arbitrary-norm separating plane , 1999, Oper. Res. Lett..

[3] Dustin Boswell,et al. Introduction to Support Vector Machines , 2002 .

[4] Bernhard Schölkopf,et al. New Support Vector Algorithms , 2000, Neural Computation.

[5] Kazushi Ikeda,et al. An asymptotic statistical analysis of support vector machines with soft margins , 2005, Neural Networks.

[6] Vladimir N. Vapnik,et al. The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[7] O. Bousquet,et al. Kernel methods and their potential use in signal processing , 2004, IEEE Signal Processing Magazine.

[8] David J. Crisp,et al. A Geometric Interpretation of v-SVM Classifiers , 1999, NIPS.

[9] David J. Crisp,et al. A Geometric Interpretation of ?-SVM Classifiers , 1999, NIPS 2000.

[10] Kristin P. Bennett,et al. Duality and Geometry in SVM Classifiers , 2000, ICML.

[12] Kazushi Ikeda. An Asymptotic Statistical Theory of Polynomial Kernel Methods , 2004, Neural Computation.