A learning algorithm for improving the classification speed of support vector machines

Jun Guo, Norikazu Takahashi, Tetsuo Nishi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

14 Citations (Scopus)

Abstract

A novel method for training support vector machines (SVMs) is proposed to speed up the SVMs in test phase. It has three main steps. First, an SVM is trained on all the training samples, thereby producing a number of support vectors. Second, the support vectors, which contribute less to the shape of the decision surface, are excluded from the training set. Finally, the SVM is re-trained only on the remaining samples. Compared to the initially trained SVM, the efficiency of the finally trained SVM is highly improved, without system degradation.

Original languageEnglish
Title of host publicationProceedings of the 2005 European Conference on Circuit Theory and Design
Pages381-384
Number of pages4
Volume3
DOIs
Publication statusPublished - 2005
Externally publishedYes
Event2005 European Conference on Circuit Theory and Design - Cork, Ireland
Duration: Aug 28 2005Sep 2 2005

Other

Other2005 European Conference on Circuit Theory and Design
CountryIreland
CityCork
Period8/28/059/2/05

Fingerprint

Learning algorithms
Support vector machines
Degradation

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Guo, J., Takahashi, N., & Nishi, T. (2005). A learning algorithm for improving the classification speed of support vector machines. In Proceedings of the 2005 European Conference on Circuit Theory and Design (Vol. 3, pp. 381-384). [1523140] https://doi.org/10.1109/ECCTD.2005.1523140

A learning algorithm for improving the classification speed of support vector machines. / Guo, Jun; Takahashi, Norikazu; Nishi, Tetsuo.

Proceedings of the 2005 European Conference on Circuit Theory and Design. Vol. 3 2005. p. 381-384 1523140.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Guo, J, Takahashi, N & Nishi, T 2005, A learning algorithm for improving the classification speed of support vector machines. in Proceedings of the 2005 European Conference on Circuit Theory and Design. vol. 3, 1523140, pp. 381-384, 2005 European Conference on Circuit Theory and Design, Cork, Ireland, 8/28/05. https://doi.org/10.1109/ECCTD.2005.1523140
Guo J, Takahashi N, Nishi T. A learning algorithm for improving the classification speed of support vector machines. In Proceedings of the 2005 European Conference on Circuit Theory and Design. Vol. 3. 2005. p. 381-384. 1523140 https://doi.org/10.1109/ECCTD.2005.1523140
Guo, Jun ; Takahashi, Norikazu ; Nishi, Tetsuo. / A learning algorithm for improving the classification speed of support vector machines. Proceedings of the 2005 European Conference on Circuit Theory and Design. Vol. 3 2005. pp. 381-384
@inproceedings{208316a9191b41bbb59b57653e14dd12,
title = "A learning algorithm for improving the classification speed of support vector machines",
abstract = "A novel method for training support vector machines (SVMs) is proposed to speed up the SVMs in test phase. It has three main steps. First, an SVM is trained on all the training samples, thereby producing a number of support vectors. Second, the support vectors, which contribute less to the shape of the decision surface, are excluded from the training set. Finally, the SVM is re-trained only on the remaining samples. Compared to the initially trained SVM, the efficiency of the finally trained SVM is highly improved, without system degradation.",
author = "Jun Guo and Norikazu Takahashi and Tetsuo Nishi",
year = "2005",
doi = "10.1109/ECCTD.2005.1523140",
language = "English",
isbn = "0780390660",
volume = "3",
pages = "381--384",
booktitle = "Proceedings of the 2005 European Conference on Circuit Theory and Design",

}

TY - GEN

T1 - A learning algorithm for improving the classification speed of support vector machines

AU - Guo, Jun

AU - Takahashi, Norikazu

AU - Nishi, Tetsuo

PY - 2005

Y1 - 2005

N2 - A novel method for training support vector machines (SVMs) is proposed to speed up the SVMs in test phase. It has three main steps. First, an SVM is trained on all the training samples, thereby producing a number of support vectors. Second, the support vectors, which contribute less to the shape of the decision surface, are excluded from the training set. Finally, the SVM is re-trained only on the remaining samples. Compared to the initially trained SVM, the efficiency of the finally trained SVM is highly improved, without system degradation.

AB - A novel method for training support vector machines (SVMs) is proposed to speed up the SVMs in test phase. It has three main steps. First, an SVM is trained on all the training samples, thereby producing a number of support vectors. Second, the support vectors, which contribute less to the shape of the decision surface, are excluded from the training set. Finally, the SVM is re-trained only on the remaining samples. Compared to the initially trained SVM, the efficiency of the finally trained SVM is highly improved, without system degradation.

UR - http://www.scopus.com/inward/record.url?scp=33748997961&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33748997961&partnerID=8YFLogxK

U2 - 10.1109/ECCTD.2005.1523140

DO - 10.1109/ECCTD.2005.1523140

M3 - Conference contribution

AN - SCOPUS:33748997961

SN - 0780390660

SN - 9780780390669

VL - 3

SP - 381

EP - 384

BT - Proceedings of the 2005 European Conference on Circuit Theory and Design

ER -