Empirical comparison of Multi-Label classification algorithms

Clifford A. Tawiah; Victor S. Sheng

Empirical comparison of Multi-Label classification algorithms

Clifford A. Tawiah, Victor S. Sheng

Computer Science

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

10 Scopus citations

Abstract

Multi-label classifications exist in many real world applications. This paper empirically studies the performance of a variety of multi-label classification algorithms. Some of them are developed based on problem transformation. Some of them are developed based on adaption. Our experimental results show that the adaptive Multi-Label K-Nearest Neighbor performs the best, followed by Random k-Label Set, followed by Classifier Chain and Binary Relevance. Adaboost. MH performs the worst, followed by Pruned Problem Transformation. Our experimental results also provide us the confidence of the correlations among multilabels. These insights shed light for future research directions on multi-label classifications.

Original language	English
Title of host publication	Proceedings of the 27th AAAI Conference on Artificial Intelligence, AAAI 2013
Pages	1645-1646
Number of pages	2
State	Published - 2013
Event	27th AAAI Conference on Artificial Intelligence, AAAI 2013 - Bellevue, WA, United States Duration: Jul 14 2013 → Jul 18 2013

Publication series

Name	Proceedings of the 27th AAAI Conference on Artificial Intelligence, AAAI 2013

Conference

Conference	27th AAAI Conference on Artificial Intelligence, AAAI 2013
Country	United States
City	Bellevue, WA
Period	07/14/13 → 07/18/13

Access to Document

Link to publication in Scopus

Fingerprint Dive into the research topics of 'Empirical comparison of Multi-Label classification algorithms'. Together they form a unique fingerprint.

Cite this

@inproceedings{292f730df1f44edc9aa056fe9f7883d8,

title = "Empirical comparison of Multi-Label classification algorithms",

abstract = "Multi-label classifications exist in many real world applications. This paper empirically studies the performance of a variety of multi-label classification algorithms. Some of them are developed based on problem transformation. Some of them are developed based on adaption. Our experimental results show that the adaptive Multi-Label K-Nearest Neighbor performs the best, followed by Random k-Label Set, followed by Classifier Chain and Binary Relevance. Adaboost. MH performs the worst, followed by Pruned Problem Transformation. Our experimental results also provide us the confidence of the correlations among multilabels. These insights shed light for future research directions on multi-label classifications.",

author = "Tawiah, {Clifford A.} and Sheng, {Victor S.}",

year = "2013",

language = "English",

isbn = "9781577356158",

series = "Proceedings of the 27th AAAI Conference on Artificial Intelligence, AAAI 2013",

pages = "1645--1646",

booktitle = "Proceedings of the 27th AAAI Conference on Artificial Intelligence, AAAI 2013",

}

Empirical comparison of Multi-Label classification algorithms. / Tawiah, Clifford A.; Sheng, Victor S.

Proceedings of the 27th AAAI Conference on Artificial Intelligence, AAAI 2013. 2013. p. 1645-1646 (Proceedings of the 27th AAAI Conference on Artificial Intelligence, AAAI 2013).

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

TY - GEN

T1 - Empirical comparison of Multi-Label classification algorithms

AU - Tawiah, Clifford A.

AU - Sheng, Victor S.

PY - 2013

Y1 - 2013

N2 - Multi-label classifications exist in many real world applications. This paper empirically studies the performance of a variety of multi-label classification algorithms. Some of them are developed based on problem transformation. Some of them are developed based on adaption. Our experimental results show that the adaptive Multi-Label K-Nearest Neighbor performs the best, followed by Random k-Label Set, followed by Classifier Chain and Binary Relevance. Adaboost. MH performs the worst, followed by Pruned Problem Transformation. Our experimental results also provide us the confidence of the correlations among multilabels. These insights shed light for future research directions on multi-label classifications.

AB - Multi-label classifications exist in many real world applications. This paper empirically studies the performance of a variety of multi-label classification algorithms. Some of them are developed based on problem transformation. Some of them are developed based on adaption. Our experimental results show that the adaptive Multi-Label K-Nearest Neighbor performs the best, followed by Random k-Label Set, followed by Classifier Chain and Binary Relevance. Adaboost. MH performs the worst, followed by Pruned Problem Transformation. Our experimental results also provide us the confidence of the correlations among multilabels. These insights shed light for future research directions on multi-label classifications.

UR - http://www.scopus.com/inward/record.url?scp=84891610061&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84891610061

SN - 9781577356158

T3 - Proceedings of the 27th AAAI Conference on Artificial Intelligence, AAAI 2013

SP - 1645

EP - 1646

BT - Proceedings of the 27th AAAI Conference on Artificial Intelligence, AAAI 2013

Y2 - 14 July 2013 through 18 July 2013

ER -