Repulsive-SVDD Classification

Phuoc NGUYEN, Dat TRAN

Research output: A Conference proceeding or a Chapter in BookConference contribution

2 Citations (Scopus)

Abstract

Support vector data description (SVDD) is a well-known kernel method that constructs a minimal hypersphere regarded as a data description for a given data set. However SVDD does not take into account any statistical distribution of the data set in constructing that optimal hypersphere, and SVDD is applied to solving one-class classification problems only. This paper proposes a new approach to SVDD to address those limitations. We formulate an optimisation problem for binary classification in which we construct two hyperspheres, one enclosing positive samples and the other enclosing negative samples, and during the optimisation process we move the two hyperspheres apart to maximise the margin between them while the data samples of each class are still inside their own hyperspheres. Experimental results show good performance for the proposed method.
Original languageEnglish
Title of host publicationAdvances in Knowledge Discovery and Data Mining
Subtitle of host publicationPAKDD 2015
EditorsTu-Bao Ho, Hiroshi Motoda, Hiroshi Motoda, Ee-Peng Lim, Tru Cao, David Cheung, Zhi-Hua Zhou
Place of PublicationCham, Switzerland
PublisherSpringer
Pages277-288
Number of pages12
Volume2
ISBN (Electronic)9783319180380
ISBN (Print)9783319180373
DOIs
Publication statusPublished - 2015
Event19th Pacific-Asia Conference on Knowledge Discovery and Data Mining - Rex hotel, Ho Chi Minh City, Viet Nam
Duration: 19 May 201522 May 2015
http://pakddsc.webfactional.com/archive/pakdd2015/

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9077
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference19th Pacific-Asia Conference on Knowledge Discovery and Data Mining
Abbreviated titlePAKDD
CountryViet Nam
CityHo Chi Minh City
Period19/05/1522/05/15
Internet address

Fingerprint

Data description

Cite this

NGUYEN, P., & TRAN, D. (2015). Repulsive-SVDD Classification. In T-B. Ho, H. Motoda, H. Motoda, E-P. Lim, T. Cao, D. Cheung, & Z-H. Zhou (Eds.), Advances in Knowledge Discovery and Data Mining: PAKDD 2015 (Vol. 2, pp. 277-288). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9077). Cham, Switzerland: Springer. https://doi.org/10.1007/978-3-319-18038-0_22
NGUYEN, Phuoc ; TRAN, Dat. / Repulsive-SVDD Classification. Advances in Knowledge Discovery and Data Mining: PAKDD 2015. editor / Tu-Bao Ho ; Hiroshi Motoda ; Hiroshi Motoda ; Ee-Peng Lim ; Tru Cao ; David Cheung ; Zhi-Hua Zhou. Vol. 2 Cham, Switzerland : Springer, 2015. pp. 277-288 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{39220302cfbe40c3ab7b11811e9a8014,
title = "Repulsive-SVDD Classification",
abstract = "Support vector data description (SVDD) is a well-known kernel method that constructs a minimal hypersphere regarded as a data description for a given data set. However SVDD does not take into account any statistical distribution of the data set in constructing that optimal hypersphere, and SVDD is applied to solving one-class classification problems only. This paper proposes a new approach to SVDD to address those limitations. We formulate an optimisation problem for binary classification in which we construct two hyperspheres, one enclosing positive samples and the other enclosing negative samples, and during the optimisation process we move the two hyperspheres apart to maximise the margin between them while the data samples of each class are still inside their own hyperspheres. Experimental results show good performance for the proposed method.",
keywords = "Classification, Repulsive SVDD, Support vector data description, Support vector machine",
author = "Phuoc NGUYEN and Dat TRAN",
year = "2015",
doi = "10.1007/978-3-319-18038-0_22",
language = "English",
isbn = "9783319180373",
volume = "2",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "277--288",
editor = "Tu-Bao Ho and Hiroshi Motoda and Hiroshi Motoda and Ee-Peng Lim and Tru Cao and David Cheung and Zhi-Hua Zhou",
booktitle = "Advances in Knowledge Discovery and Data Mining",
address = "Netherlands",

}

NGUYEN, P & TRAN, D 2015, Repulsive-SVDD Classification. in T-B Ho, H Motoda, H Motoda, E-P Lim, T Cao, D Cheung & Z-H Zhou (eds), Advances in Knowledge Discovery and Data Mining: PAKDD 2015. vol. 2, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9077, Springer, Cham, Switzerland, pp. 277-288, 19th Pacific-Asia Conference on Knowledge Discovery and Data Mining, Ho Chi Minh City, Viet Nam, 19/05/15. https://doi.org/10.1007/978-3-319-18038-0_22

Repulsive-SVDD Classification. / NGUYEN, Phuoc; TRAN, Dat.

Advances in Knowledge Discovery and Data Mining: PAKDD 2015. ed. / Tu-Bao Ho; Hiroshi Motoda; Hiroshi Motoda; Ee-Peng Lim; Tru Cao; David Cheung; Zhi-Hua Zhou. Vol. 2 Cham, Switzerland : Springer, 2015. p. 277-288 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9077).

Research output: A Conference proceeding or a Chapter in BookConference contribution

TY - GEN

T1 - Repulsive-SVDD Classification

AU - NGUYEN, Phuoc

AU - TRAN, Dat

PY - 2015

Y1 - 2015

N2 - Support vector data description (SVDD) is a well-known kernel method that constructs a minimal hypersphere regarded as a data description for a given data set. However SVDD does not take into account any statistical distribution of the data set in constructing that optimal hypersphere, and SVDD is applied to solving one-class classification problems only. This paper proposes a new approach to SVDD to address those limitations. We formulate an optimisation problem for binary classification in which we construct two hyperspheres, one enclosing positive samples and the other enclosing negative samples, and during the optimisation process we move the two hyperspheres apart to maximise the margin between them while the data samples of each class are still inside their own hyperspheres. Experimental results show good performance for the proposed method.

AB - Support vector data description (SVDD) is a well-known kernel method that constructs a minimal hypersphere regarded as a data description for a given data set. However SVDD does not take into account any statistical distribution of the data set in constructing that optimal hypersphere, and SVDD is applied to solving one-class classification problems only. This paper proposes a new approach to SVDD to address those limitations. We formulate an optimisation problem for binary classification in which we construct two hyperspheres, one enclosing positive samples and the other enclosing negative samples, and during the optimisation process we move the two hyperspheres apart to maximise the margin between them while the data samples of each class are still inside their own hyperspheres. Experimental results show good performance for the proposed method.

KW - Classification

KW - Repulsive SVDD

KW - Support vector data description

KW - Support vector machine

UR - http://www.scopus.com/inward/record.url?scp=84945925650&partnerID=8YFLogxK

UR - http://www.mendeley.com/research/repulsivesvdd-classification

U2 - 10.1007/978-3-319-18038-0_22

DO - 10.1007/978-3-319-18038-0_22

M3 - Conference contribution

SN - 9783319180373

VL - 2

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

SP - 277

EP - 288

BT - Advances in Knowledge Discovery and Data Mining

A2 - Ho, Tu-Bao

A2 - Motoda, Hiroshi

A2 - Motoda, Hiroshi

A2 - Lim, Ee-Peng

A2 - Cao, Tru

A2 - Cheung, David

A2 - Zhou, Zhi-Hua

PB - Springer

CY - Cham, Switzerland

ER -

NGUYEN P, TRAN D. Repulsive-SVDD Classification. In Ho T-B, Motoda H, Motoda H, Lim E-P, Cao T, Cheung D, Zhou Z-H, editors, Advances in Knowledge Discovery and Data Mining: PAKDD 2015. Vol. 2. Cham, Switzerland: Springer. 2015. p. 277-288. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-18038-0_22