Ensemble approaches of support vector machines for multiclass classification

Jun Ki Min, Jin Hyuk Hong, Sung Bae Cho

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

2 Citations (Scopus)

Abstract

Support vector machine (SVM) which was originally designed for binary classification has achieved superior performance in various classification problems. In order to extend it to multiclass classification, one popular approach is to consider the problem as a collection of binary classification problems. Majority voting or winner-takes-all is then applied to combine those outputs, but it often causes problems to consider tie-breaks and tune the weights of individual classifiers. This paper presents two novel ensemble approaches: probabilistic ordering of one-vs-rest (OVR) SVMs with naïve Bayes classifier and multiple decision templates of OVR SVMs. Experiments with multiclass datasets have shown the usefulness of the ensemble methods.

Original languageEnglish
Title of host publicationPattern Recognition and Machine Intelligence - Second International Conference, PReMI 2007, Proceedings
Pages1-10
Number of pages10
Publication statusPublished - 2007 Dec 1
Event2nd International Conference on Pattern Recognition and Machine Intelligence, PReMI 2007 - Kolkata, India
Duration: 2007 Dec 182007 Dec 22

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4815 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other2nd International Conference on Pattern Recognition and Machine Intelligence, PReMI 2007
CountryIndia
CityKolkata
Period07/12/1807/12/22

Fingerprint

Multi-class Classification
Binary Classification
Classification Problems
Support vector machines
Support Vector Machine
Ensemble
Bayes Classifier
Winner-take-all
Majority Voting
Ensemble Methods
Probabilistic Approach
Multi-class
Tie
Template
Classifiers
Classifier
Output
Experiment
Experiments

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Min, J. K., Hong, J. H., & Cho, S. B. (2007). Ensemble approaches of support vector machines for multiclass classification. In Pattern Recognition and Machine Intelligence - Second International Conference, PReMI 2007, Proceedings (pp. 1-10). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4815 LNCS).
Min, Jun Ki ; Hong, Jin Hyuk ; Cho, Sung Bae. / Ensemble approaches of support vector machines for multiclass classification. Pattern Recognition and Machine Intelligence - Second International Conference, PReMI 2007, Proceedings. 2007. pp. 1-10 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{f175ce25578845d2b01969474dfe5a79,
title = "Ensemble approaches of support vector machines for multiclass classification",
abstract = "Support vector machine (SVM) which was originally designed for binary classification has achieved superior performance in various classification problems. In order to extend it to multiclass classification, one popular approach is to consider the problem as a collection of binary classification problems. Majority voting or winner-takes-all is then applied to combine those outputs, but it often causes problems to consider tie-breaks and tune the weights of individual classifiers. This paper presents two novel ensemble approaches: probabilistic ordering of one-vs-rest (OVR) SVMs with na{\"i}ve Bayes classifier and multiple decision templates of OVR SVMs. Experiments with multiclass datasets have shown the usefulness of the ensemble methods.",
author = "Min, {Jun Ki} and Hong, {Jin Hyuk} and Cho, {Sung Bae}",
year = "2007",
month = "12",
day = "1",
language = "English",
isbn = "3540770453",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "1--10",
booktitle = "Pattern Recognition and Machine Intelligence - Second International Conference, PReMI 2007, Proceedings",

}

Min, JK, Hong, JH & Cho, SB 2007, Ensemble approaches of support vector machines for multiclass classification. in Pattern Recognition and Machine Intelligence - Second International Conference, PReMI 2007, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4815 LNCS, pp. 1-10, 2nd International Conference on Pattern Recognition and Machine Intelligence, PReMI 2007, Kolkata, India, 07/12/18.

Ensemble approaches of support vector machines for multiclass classification. / Min, Jun Ki; Hong, Jin Hyuk; Cho, Sung Bae.

Pattern Recognition and Machine Intelligence - Second International Conference, PReMI 2007, Proceedings. 2007. p. 1-10 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4815 LNCS).

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

TY - GEN

T1 - Ensemble approaches of support vector machines for multiclass classification

AU - Min, Jun Ki

AU - Hong, Jin Hyuk

AU - Cho, Sung Bae

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Support vector machine (SVM) which was originally designed for binary classification has achieved superior performance in various classification problems. In order to extend it to multiclass classification, one popular approach is to consider the problem as a collection of binary classification problems. Majority voting or winner-takes-all is then applied to combine those outputs, but it often causes problems to consider tie-breaks and tune the weights of individual classifiers. This paper presents two novel ensemble approaches: probabilistic ordering of one-vs-rest (OVR) SVMs with naïve Bayes classifier and multiple decision templates of OVR SVMs. Experiments with multiclass datasets have shown the usefulness of the ensemble methods.

AB - Support vector machine (SVM) which was originally designed for binary classification has achieved superior performance in various classification problems. In order to extend it to multiclass classification, one popular approach is to consider the problem as a collection of binary classification problems. Majority voting or winner-takes-all is then applied to combine those outputs, but it often causes problems to consider tie-breaks and tune the weights of individual classifiers. This paper presents two novel ensemble approaches: probabilistic ordering of one-vs-rest (OVR) SVMs with naïve Bayes classifier and multiple decision templates of OVR SVMs. Experiments with multiclass datasets have shown the usefulness of the ensemble methods.

UR - http://www.scopus.com/inward/record.url?scp=38149074049&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=38149074049&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:38149074049

SN - 3540770453

SN - 9783540770459

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 1

EP - 10

BT - Pattern Recognition and Machine Intelligence - Second International Conference, PReMI 2007, Proceedings

ER -

Min JK, Hong JH, Cho SB. Ensemble approaches of support vector machines for multiclass classification. In Pattern Recognition and Machine Intelligence - Second International Conference, PReMI 2007, Proceedings. 2007. p. 1-10. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).