Fast leave-one-out evaluation and improvement on inference for LS-SVMs

Author

Ying, Zhao ; Keong, Kwoh Chee

Author_Institution

Bioinf. Res. Centre, Nanyang Technol. Univ., Singapore

Volume

3

fYear

2004

fDate

23-26 Aug. 2004

Firstpage

494

Abstract

In this paper, a fast leave-one-out (LOO) evaluation formula is introduced for least squares support vector machine (LS-SVM) classifiers. The computation cost can be reduced to approximately 1/N when compared to normal LOO procedure (N is the number of training samples). Inspired by its fast speed, we are able to use it to replace the original level 3 posterior probability approximation formula of the Bayesian framework for LS-SVM classifiers. The improved inference framework shows higher generalization performance and faster computation speed.

Keywords

Bayes methods; least squares approximations; pattern classification; probability; sampling methods; support vector machines; Bayesian method; SVM classifiers; fast leave-one-out evaluation; least squares classifier; posterior probability approximation; sampling method; support vector machine classifier; support vector machine inference; Bayesian methods; Bioinformatics; Computational efficiency; High performance computing; Lagrangian functions; Least squares approximation; Least squares methods; Support vector machine classification; Support vector machines; System testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Pattern Recognition, 2004. ICPR 2004. Proceedings of the 17th International Conference on

ISSN

1051-4651

Print_ISBN

0-7695-2128-2

Type

conf

DOI

10.1109/ICPR.2004.1334574

Filename

1334574