Improving multi-label classification performance by label constraints

Benhui Chen, Xuefen Hong, Lihua Duan, Takayuki Furuzuki

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

6 Citations (Scopus)

Abstract

Multi-label classification is an extension of traditional classification problem in which each instance is associated with a set of labels. For some multi-label classification tasks, labels are usually overlapped and correlated, and some implicit constraint rules are existed among the labels. This paper presents an improved multi-label classification method based on label ranking strategy and label constraints. Firstly, one-against-all decomposition technique is used to divide a multilabel classification task into multiple independent binary classification sub-problems. One binary SVM classifier is trained for each label. Secondly, based on training data, label constraint rules are mined by association rule learning method. Thirdly, a correction model based on label constraints is used to correct the probabilistic outputs of SVM classifiers for label ranking. Experiment results on three well-known multi-label benchmark datasets show that the proposed method outperforms some conventional multi-label classification methods.

Original languageEnglish
Title of host publicationProceedings of the International Joint Conference on Neural Networks
DOIs
Publication statusPublished - 2013
Event2013 International Joint Conference on Neural Networks, IJCNN 2013 - Dallas, TX
Duration: 2013 Aug 42013 Aug 9

Other

Other2013 International Joint Conference on Neural Networks, IJCNN 2013
CityDallas, TX
Period13/8/413/8/9

Fingerprint

Labels
Classifiers
Association rules
Decomposition

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence

Cite this

Chen, B., Hong, X., Duan, L., & Furuzuki, T. (2013). Improving multi-label classification performance by label constraints. In Proceedings of the International Joint Conference on Neural Networks [6706861] https://doi.org/10.1109/IJCNN.2013.6706861

Improving multi-label classification performance by label constraints. / Chen, Benhui; Hong, Xuefen; Duan, Lihua; Furuzuki, Takayuki.

Proceedings of the International Joint Conference on Neural Networks. 2013. 6706861.

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

Chen, B, Hong, X, Duan, L & Furuzuki, T 2013, Improving multi-label classification performance by label constraints. in Proceedings of the International Joint Conference on Neural Networks., 6706861, 2013 International Joint Conference on Neural Networks, IJCNN 2013, Dallas, TX, 13/8/4. https://doi.org/10.1109/IJCNN.2013.6706861
Chen B, Hong X, Duan L, Furuzuki T. Improving multi-label classification performance by label constraints. In Proceedings of the International Joint Conference on Neural Networks. 2013. 6706861 https://doi.org/10.1109/IJCNN.2013.6706861
Chen, Benhui ; Hong, Xuefen ; Duan, Lihua ; Furuzuki, Takayuki. / Improving multi-label classification performance by label constraints. Proceedings of the International Joint Conference on Neural Networks. 2013.
@inproceedings{42613015e8ad4828b40552229c2d21dc,
title = "Improving multi-label classification performance by label constraints",
abstract = "Multi-label classification is an extension of traditional classification problem in which each instance is associated with a set of labels. For some multi-label classification tasks, labels are usually overlapped and correlated, and some implicit constraint rules are existed among the labels. This paper presents an improved multi-label classification method based on label ranking strategy and label constraints. Firstly, one-against-all decomposition technique is used to divide a multilabel classification task into multiple independent binary classification sub-problems. One binary SVM classifier is trained for each label. Secondly, based on training data, label constraint rules are mined by association rule learning method. Thirdly, a correction model based on label constraints is used to correct the probabilistic outputs of SVM classifiers for label ranking. Experiment results on three well-known multi-label benchmark datasets show that the proposed method outperforms some conventional multi-label classification methods.",
author = "Benhui Chen and Xuefen Hong and Lihua Duan and Takayuki Furuzuki",
year = "2013",
doi = "10.1109/IJCNN.2013.6706861",
language = "English",
isbn = "9781467361293",
booktitle = "Proceedings of the International Joint Conference on Neural Networks",

}

TY - GEN

T1 - Improving multi-label classification performance by label constraints

AU - Chen, Benhui

AU - Hong, Xuefen

AU - Duan, Lihua

AU - Furuzuki, Takayuki

PY - 2013

Y1 - 2013

N2 - Multi-label classification is an extension of traditional classification problem in which each instance is associated with a set of labels. For some multi-label classification tasks, labels are usually overlapped and correlated, and some implicit constraint rules are existed among the labels. This paper presents an improved multi-label classification method based on label ranking strategy and label constraints. Firstly, one-against-all decomposition technique is used to divide a multilabel classification task into multiple independent binary classification sub-problems. One binary SVM classifier is trained for each label. Secondly, based on training data, label constraint rules are mined by association rule learning method. Thirdly, a correction model based on label constraints is used to correct the probabilistic outputs of SVM classifiers for label ranking. Experiment results on three well-known multi-label benchmark datasets show that the proposed method outperforms some conventional multi-label classification methods.

AB - Multi-label classification is an extension of traditional classification problem in which each instance is associated with a set of labels. For some multi-label classification tasks, labels are usually overlapped and correlated, and some implicit constraint rules are existed among the labels. This paper presents an improved multi-label classification method based on label ranking strategy and label constraints. Firstly, one-against-all decomposition technique is used to divide a multilabel classification task into multiple independent binary classification sub-problems. One binary SVM classifier is trained for each label. Secondly, based on training data, label constraint rules are mined by association rule learning method. Thirdly, a correction model based on label constraints is used to correct the probabilistic outputs of SVM classifiers for label ranking. Experiment results on three well-known multi-label benchmark datasets show that the proposed method outperforms some conventional multi-label classification methods.

UR - http://www.scopus.com/inward/record.url?scp=84893618184&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84893618184&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2013.6706861

DO - 10.1109/IJCNN.2013.6706861

M3 - Conference contribution

AN - SCOPUS:84893618184

SN - 9781467361293

BT - Proceedings of the International Joint Conference on Neural Networks

ER -