Pegasos Algorithm for One-Class Support Vector Machine

Changki LEE

doi:10.1587/transinf.E96.D.1223

Pegasos Algorithm for One-Class Support Vector Machine

Changki LEE

Full Text Views

0

Share
Cite this

Summary :

Training one-class support vector machines (one-class SVMs) involves solving a quadratic programming (QP) problem. By increasing the number of training samples, solving this QP problem becomes intractable. In this paper, we describe a modified Pegasos algorithm for fast training of one-class SVMs. We show that this algorithm is much faster than the standard one-class SVM without loss of performance in the case of linear kernel.

Publication: IEICE TRANSACTIONS on Information Vol.E96-D No.5 pp.1223-1226

Publication Date: 2013/05/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.E96.D.1223

Type of Manuscript: LETTER

Category: Artificial Intelligence, Data Mining

Cite this

Copy

Changki LEE, "Pegasos Algorithm for One-Class Support Vector Machine" in IEICE TRANSACTIONS on Information, vol. E96-D, no. 5, pp. 1223-1226, May 2013, doi: 10.1587/transinf.E96.D.1223.
Abstract: Training one-class support vector machines (one-class SVMs) involves solving a quadratic programming (QP) problem. By increasing the number of training samples, solving this QP problem becomes intractable. In this paper, we describe a modified Pegasos algorithm for fast training of one-class SVMs. We show that this algorithm is much faster than the standard one-class SVM without loss of performance in the case of linear kernel.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.E96.D.1223/_p

Copy

@ARTICLE{e96-d_5_1223,
author={Changki LEE, },
journal={IEICE TRANSACTIONS on Information},
title={Pegasos Algorithm for One-Class Support Vector Machine},
year={2013},
volume={E96-D},
number={5},
pages={1223-1226},
abstract={Training one-class support vector machines (one-class SVMs) involves solving a quadratic programming (QP) problem. By increasing the number of training samples, solving this QP problem becomes intractable. In this paper, we describe a modified Pegasos algorithm for fast training of one-class SVMs. We show that this algorithm is much faster than the standard one-class SVM without loss of performance in the case of linear kernel.},
keywords={},
doi={10.1587/transinf.E96.D.1223},
ISSN={1745-1361},
month={May},}

Copy

TY - JOUR
TI - Pegasos Algorithm for One-Class Support Vector Machine
T2 - IEICE TRANSACTIONS on Information
SP - 1223
EP - 1226
AU - Changki LEE
PY - 2013
DO - 10.1587/transinf.E96.D.1223
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E96-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2013
AB - Training one-class support vector machines (one-class SVMs) involves solving a quadratic programming (QP) problem. By increasing the number of training samples, solving this QP problem becomes intractable. In this paper, we describe a modified Pegasos algorithm for fast training of one-class SVMs. We show that this algorithm is much faster than the standard one-class SVM without loss of performance in the case of linear kernel.
ER -