Evidential multi-label classification approach to learning from data with imprecise labels

Zoulficar Younes; Fahed Abdallah; Thierry Denœux

doi:10.1007/978-3-642-14049-5_13

Evidential multi-label classification approach to learning from data with imprecise labels

Zoulficar Younes, Fahed Abdallah, Thierry Denœux

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

10 Citations (Scopus)

Abstract

Multi-label classification problems arise in many real-world applications. Classically, in order to construct a multi-label classifier, we assume the existence of a labeled training set, where each instance is associated with a set of labels, and the task is to output a label set for each unseen instance. However, it is not always possible to have perfectly labeled data. In many problems, there is no ground truth for assigning unambiguously a label set to each instance, and several experts have to be consulted. Due to conflicts and lack of knowledge, labels might be wrongly assigned to some instances. This paper describes an evidence formalism suitable to study multi-label classification problems where the training datasets are imperfectly labelled. Several applications demonstrate the efficiency of our apporach.

Original language	English
Title of host publication	Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings
Pages	119-128
Number of pages	10
DOIs	https://doi.org/10.1007/978-3-642-14049-5_13
Publication status	Published - 2010
Externally published	Yes
Event	13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010 - Dortmund, Germany Duration: 28 Jun 2010 → 2 Jul 2010

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	6178 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010
Country/Territory	Germany
City	Dortmund
Period	28/06/10 → 2/07/10

Access to Document

10.1007/978-3-642-14049-5_13

Cite this

Younes, Z., Abdallah, F., & Denœux, T. (2010). Evidential multi-label classification approach to learning from data with imprecise labels. In Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings (pp. 119-128). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6178 LNAI). https://doi.org/10.1007/978-3-642-14049-5_13

Younes, Zoulficar ; Abdallah, Fahed ; Denœux, Thierry. / Evidential multi-label classification approach to learning from data with imprecise labels. Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings. 2010. pp. 119-128 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{c93ee50061dd438995a0da337411c688,

title = "Evidential multi-label classification approach to learning from data with imprecise labels",

abstract = "Multi-label classification problems arise in many real-world applications. Classically, in order to construct a multi-label classifier, we assume the existence of a labeled training set, where each instance is associated with a set of labels, and the task is to output a label set for each unseen instance. However, it is not always possible to have perfectly labeled data. In many problems, there is no ground truth for assigning unambiguously a label set to each instance, and several experts have to be consulted. Due to conflicts and lack of knowledge, labels might be wrongly assigned to some instances. This paper describes an evidence formalism suitable to study multi-label classification problems where the training datasets are imperfectly labelled. Several applications demonstrate the efficiency of our apporach.",

author = "Zoulficar Younes and Fahed Abdallah and Thierry Den{\oe}ux",

year = "2010",

doi = "10.1007/978-3-642-14049-5_13",

language = "English",

isbn = "3642140483",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "119--128",

booktitle = "Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings",

note = "13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010 ; Conference date: 28-06-2010 Through 02-07-2010",

}

Younes, Z, Abdallah, F & Denœux, T 2010, Evidential multi-label classification approach to learning from data with imprecise labels. in Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 6178 LNAI, pp. 119-128, 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Dortmund, Germany, 28/06/10. https://doi.org/10.1007/978-3-642-14049-5_13

Evidential multi-label classification approach to learning from data with imprecise labels. / Younes, Zoulficar; Abdallah, Fahed; Denœux, Thierry.
Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings. 2010. p. 119-128 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6178 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Evidential multi-label classification approach to learning from data with imprecise labels

AU - Younes, Zoulficar

AU - Abdallah, Fahed

AU - Denœux, Thierry

PY - 2010

Y1 - 2010

N2 - Multi-label classification problems arise in many real-world applications. Classically, in order to construct a multi-label classifier, we assume the existence of a labeled training set, where each instance is associated with a set of labels, and the task is to output a label set for each unseen instance. However, it is not always possible to have perfectly labeled data. In many problems, there is no ground truth for assigning unambiguously a label set to each instance, and several experts have to be consulted. Due to conflicts and lack of knowledge, labels might be wrongly assigned to some instances. This paper describes an evidence formalism suitable to study multi-label classification problems where the training datasets are imperfectly labelled. Several applications demonstrate the efficiency of our apporach.

AB - Multi-label classification problems arise in many real-world applications. Classically, in order to construct a multi-label classifier, we assume the existence of a labeled training set, where each instance is associated with a set of labels, and the task is to output a label set for each unseen instance. However, it is not always possible to have perfectly labeled data. In many problems, there is no ground truth for assigning unambiguously a label set to each instance, and several experts have to be consulted. Due to conflicts and lack of knowledge, labels might be wrongly assigned to some instances. This paper describes an evidence formalism suitable to study multi-label classification problems where the training datasets are imperfectly labelled. Several applications demonstrate the efficiency of our apporach.

UR - http://www.scopus.com/inward/record.url?scp=77954867886&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-14049-5_13

DO - 10.1007/978-3-642-14049-5_13

M3 - Conference contribution

AN - SCOPUS:77954867886

SN - 3642140483

SN - 9783642140488

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

SP - 119

EP - 128

BT - Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings

T2 - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010

Y2 - 28 June 2010 through 2 July 2010

ER -

Younes Z, Abdallah F, Denœux T. Evidential multi-label classification approach to learning from data with imprecise labels. In Computational Intelligence for Knowledge-Based Systems Design - 13th International Conference on Information Processing and Management of Uncertainty, IPMU 2010, Proceedings. 2010. p. 119-128. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-642-14049-5_13

Evidential multi-label classification approach to learning from data with imprecise labels

Abstract

Publication series

Conference

Access to Document

Other files and links

Cite this