Design efficient support vector machine for fast classification

Yiqiang Zhan, Dinggang Shen

Research output: Contribution to journalArticle

62 Citations (Scopus)

Abstract

This paper presents a four-step training method for increasing the efficiency of support vector machine (SVM). First, a SVM is initially trained by all the training samples, thereby producing a number of support vectors. Second, the support vectors, which make the hypersurface highly convoluted, are excluded from the training set. Third, the SVM is re-trained only by the remaining samples in the training set, Finally, the complexity of the trained SVM is further reduced by approximating the separation hypersurface with a subset of the support vectors. Compared to the initially trained SVM by all samples, the efficiency of the finally-trained SVM is highly improved, without system degradation.

Original languageEnglish
Pages (from-to)157-161
Number of pages5
JournalPattern Recognition
Volume38
Issue number1
DOIs
Publication statusPublished - 2005 Jan 1
Externally publishedYes

Fingerprint

Support vector machines
Degradation

Keywords

  • Computational efficiency
  • Support vector machine
  • Training method

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Design efficient support vector machine for fast classification. / Zhan, Yiqiang; Shen, Dinggang.

In: Pattern Recognition, Vol. 38, No. 1, 01.01.2005, p. 157-161.

Research output: Contribution to journalArticle

@article{6be36d63ad2a41f9ac6b06594c4dcc82,
title = "Design efficient support vector machine for fast classification",
abstract = "This paper presents a four-step training method for increasing the efficiency of support vector machine (SVM). First, a SVM is initially trained by all the training samples, thereby producing a number of support vectors. Second, the support vectors, which make the hypersurface highly convoluted, are excluded from the training set. Third, the SVM is re-trained only by the remaining samples in the training set, Finally, the complexity of the trained SVM is further reduced by approximating the separation hypersurface with a subset of the support vectors. Compared to the initially trained SVM by all samples, the efficiency of the finally-trained SVM is highly improved, without system degradation.",
keywords = "Computational efficiency, Support vector machine, Training method",
author = "Yiqiang Zhan and Dinggang Shen",
year = "2005",
month = "1",
day = "1",
doi = "10.1016/j.patcog.2004.06.001",
language = "English",
volume = "38",
pages = "157--161",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",
number = "1",

}

TY - JOUR

T1 - Design efficient support vector machine for fast classification

AU - Zhan, Yiqiang

AU - Shen, Dinggang

PY - 2005/1/1

Y1 - 2005/1/1

N2 - This paper presents a four-step training method for increasing the efficiency of support vector machine (SVM). First, a SVM is initially trained by all the training samples, thereby producing a number of support vectors. Second, the support vectors, which make the hypersurface highly convoluted, are excluded from the training set. Third, the SVM is re-trained only by the remaining samples in the training set, Finally, the complexity of the trained SVM is further reduced by approximating the separation hypersurface with a subset of the support vectors. Compared to the initially trained SVM by all samples, the efficiency of the finally-trained SVM is highly improved, without system degradation.

AB - This paper presents a four-step training method for increasing the efficiency of support vector machine (SVM). First, a SVM is initially trained by all the training samples, thereby producing a number of support vectors. Second, the support vectors, which make the hypersurface highly convoluted, are excluded from the training set. Third, the SVM is re-trained only by the remaining samples in the training set, Finally, the complexity of the trained SVM is further reduced by approximating the separation hypersurface with a subset of the support vectors. Compared to the initially trained SVM by all samples, the efficiency of the finally-trained SVM is highly improved, without system degradation.

KW - Computational efficiency

KW - Support vector machine

KW - Training method

UR - http://www.scopus.com/inward/record.url?scp=4644347704&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=4644347704&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2004.06.001

DO - 10.1016/j.patcog.2004.06.001

M3 - Article

AN - SCOPUS:4644347704

VL - 38

SP - 157

EP - 161

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

IS - 1

ER -