Combining binary-SVM and pairwise label constraints for multi-label classification

Weifeng Gu, Benhui Chen, Takayuki Furuzuki

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

2 Citations (Scopus)

Abstract

Multi-label classification is an extension of traditional classification problem in which each instance is associated with a set of labels. Recent research has shown that the ranking approach is an effective way to solve this problem. In the multi-labeled sets, classes are often related to each other. Some implicit constraint rules are existed among the labels. So we present a novel multi-label ranking algorithm inspired by the pairwise constraint rules mined from the training set to enhance the existing method. In this method, one-against-all decomposition technique is used firstly to divide a multi-label problem into binary class sub-problems. A rank list is generated by combining the probabilistic outputs of each binary Support Vector Machine (SVM) classifier. Label constraint rules are learned by minimizing the ranking loss. Experimental performance evaluation on well-known multi-label benchmark datasets show that our method improves the classification accuracy efficiently, compared with some existed methods.

Original languageEnglish
Title of host publicationConference Proceedings - IEEE International Conference on Systems, Man and Cybernetics
Pages4176-4181
Number of pages6
DOIs
Publication statusPublished - 2010
Event2010 IEEE International Conference on Systems, Man and Cybernetics, SMC 2010 - Istanbul
Duration: 2010 Oct 102010 Oct 13

Other

Other2010 IEEE International Conference on Systems, Man and Cybernetics, SMC 2010
CityIstanbul
Period10/10/1010/10/13

Fingerprint

Support vector machines
Labels
Classifiers
Decomposition

Keywords

  • Constraint rules
  • Multi-label classification
  • Support vector machine

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Control and Systems Engineering
  • Human-Computer Interaction

Cite this

Gu, W., Chen, B., & Furuzuki, T. (2010). Combining binary-SVM and pairwise label constraints for multi-label classification. In Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics (pp. 4176-4181). [5642395] https://doi.org/10.1109/ICSMC.2010.5642395

Combining binary-SVM and pairwise label constraints for multi-label classification. / Gu, Weifeng; Chen, Benhui; Furuzuki, Takayuki.

Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. 2010. p. 4176-4181 5642395.

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

Gu, W, Chen, B & Furuzuki, T 2010, Combining binary-SVM and pairwise label constraints for multi-label classification. in Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics., 5642395, pp. 4176-4181, 2010 IEEE International Conference on Systems, Man and Cybernetics, SMC 2010, Istanbul, 10/10/10. https://doi.org/10.1109/ICSMC.2010.5642395
Gu W, Chen B, Furuzuki T. Combining binary-SVM and pairwise label constraints for multi-label classification. In Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. 2010. p. 4176-4181. 5642395 https://doi.org/10.1109/ICSMC.2010.5642395
Gu, Weifeng ; Chen, Benhui ; Furuzuki, Takayuki. / Combining binary-SVM and pairwise label constraints for multi-label classification. Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. 2010. pp. 4176-4181
@inproceedings{5f6cbb504906410bbee3ff820893eef9,
title = "Combining binary-SVM and pairwise label constraints for multi-label classification",
abstract = "Multi-label classification is an extension of traditional classification problem in which each instance is associated with a set of labels. Recent research has shown that the ranking approach is an effective way to solve this problem. In the multi-labeled sets, classes are often related to each other. Some implicit constraint rules are existed among the labels. So we present a novel multi-label ranking algorithm inspired by the pairwise constraint rules mined from the training set to enhance the existing method. In this method, one-against-all decomposition technique is used firstly to divide a multi-label problem into binary class sub-problems. A rank list is generated by combining the probabilistic outputs of each binary Support Vector Machine (SVM) classifier. Label constraint rules are learned by minimizing the ranking loss. Experimental performance evaluation on well-known multi-label benchmark datasets show that our method improves the classification accuracy efficiently, compared with some existed methods.",
keywords = "Constraint rules, Multi-label classification, Support vector machine",
author = "Weifeng Gu and Benhui Chen and Takayuki Furuzuki",
year = "2010",
doi = "10.1109/ICSMC.2010.5642395",
language = "English",
isbn = "9781424465880",
pages = "4176--4181",
booktitle = "Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics",

}

TY - GEN

T1 - Combining binary-SVM and pairwise label constraints for multi-label classification

AU - Gu, Weifeng

AU - Chen, Benhui

AU - Furuzuki, Takayuki

PY - 2010

Y1 - 2010

N2 - Multi-label classification is an extension of traditional classification problem in which each instance is associated with a set of labels. Recent research has shown that the ranking approach is an effective way to solve this problem. In the multi-labeled sets, classes are often related to each other. Some implicit constraint rules are existed among the labels. So we present a novel multi-label ranking algorithm inspired by the pairwise constraint rules mined from the training set to enhance the existing method. In this method, one-against-all decomposition technique is used firstly to divide a multi-label problem into binary class sub-problems. A rank list is generated by combining the probabilistic outputs of each binary Support Vector Machine (SVM) classifier. Label constraint rules are learned by minimizing the ranking loss. Experimental performance evaluation on well-known multi-label benchmark datasets show that our method improves the classification accuracy efficiently, compared with some existed methods.

AB - Multi-label classification is an extension of traditional classification problem in which each instance is associated with a set of labels. Recent research has shown that the ranking approach is an effective way to solve this problem. In the multi-labeled sets, classes are often related to each other. Some implicit constraint rules are existed among the labels. So we present a novel multi-label ranking algorithm inspired by the pairwise constraint rules mined from the training set to enhance the existing method. In this method, one-against-all decomposition technique is used firstly to divide a multi-label problem into binary class sub-problems. A rank list is generated by combining the probabilistic outputs of each binary Support Vector Machine (SVM) classifier. Label constraint rules are learned by minimizing the ranking loss. Experimental performance evaluation on well-known multi-label benchmark datasets show that our method improves the classification accuracy efficiently, compared with some existed methods.

KW - Constraint rules

KW - Multi-label classification

KW - Support vector machine

UR - http://www.scopus.com/inward/record.url?scp=78751537298&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78751537298&partnerID=8YFLogxK

U2 - 10.1109/ICSMC.2010.5642395

DO - 10.1109/ICSMC.2010.5642395

M3 - Conference contribution

SN - 9781424465880

SP - 4176

EP - 4181

BT - Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics

ER -