The generalization error of the symmetric and scaled support vector machines

Author

Feng, Jianfeng ; Williams, Peter

Author_Institution

Sch. of Cognitive & Comput. Sci., Sussex Univ., Brighton, UK

Volume

12

Issue

5

fYear

2001

fDate

9/1/2001 12:00:00 AM

Firstpage

1255

Lastpage

1260

Abstract

It is generally believed that the support vector machine (SVM) optimizes the generalization error and outperforms other learning machines. We show analytically, by concrete examples in the one dimensional case, that the SVM does improve the mean and standard deviation of the generalization error by a constant factor, compared to the worst learning machine. Our approach is in terms of the extreme value theory and both the mean and variance of the generalization errors are calculated exactly for all the cases considered. We propose a new version of the SVM , called the scaled SVM, which can further reduce the mean of the generalization error of the SVM

Keywords

error analysis; generalisation (artificial intelligence); learning (artificial intelligence); learning automata; extreme value theory; generalization error; learning machines; scaled support vector machines; Approximation error; Biomedical computing; Cardiology; Control systems; Digital filters; Electrocardiography; Feature extraction; Neural networks; Radial basis function networks; Support vector machines;

fLanguage

English

Journal_Title

Neural Networks, IEEE Transactions on

Publisher

ieee

ISSN

1045-9227

Type

jour

DOI

10.1109/72.950155

Filename

950155