Fast One-Class support Vector Machine for novelty detection

Trung Le; Dinh Phung; Khanh Nguyen; Svetha Venkatesh

doi:10.1007/978-3-319-18032-8_15

Fast One-Class support Vector Machine for novelty detection

Trung Le, Dinh Phung, Khanh Nguyen, Svetha Venkatesh

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

4 Citations (Scopus)

Abstract

Novelty detection arises as an important learning task in several applications. Kernel-based approach to novelty detection has been widely used due to its theoretical rigor and elegance of geometric interpretation. However, computational complexity is a major obstacle in this approach. In this paper, leveraging on the cutting-plane framework with the well-known One-Class Support Vector Machine, we present a new solution that can scale up seamlessly with data. The first solution is exact and linear when viewed through the cutting-plane; the second employed a sampling strategy that remarkably has a constant computational complexity defined relatively to the probability of approximation accuracy. Several datasets are benchmarked to demonstrate the credibility of our framework.

Original language	English
Title of host publication	Advances in Knowledge Discovery and Data Mining
Subtitle of host publication	19th Pacific-Asia Conference, PAKDD 2015 Ho Chi Minh City, Vietnam, May 19–22, 2015 Proceedings, Part II
Editors	Tru Cao, Ee-Peng Lim, Zhi-Hua Zhou, Tu-Bao Ho, David Cheung, Hiroshi Motoda
Place of Publication	Cham Switzerland
Publisher	Springer
Pages	189-200
Number of pages	12
ISBN (Electronic)	9783319180328
ISBN (Print)	9783319180311
DOIs	https://doi.org/10.1007/978-3-319-18032-8_15
Publication status	Published - 2015
Externally published	Yes
Event	Pacific Asia Conference on Information Systems 2015 - Singapore, Singapore Duration: 5 Jul 2015 → 9 Jul 2015 Conference number: 19th https://web.archive.org/web/20150728131139/http://www.pacis2015.org/ https://aisel.aisnet.org/pacis2015/ (Proceedings)

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	9078
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	Pacific Asia Conference on Information Systems 2015
Abbreviated title	PACIS 2015
Country/Territory	Singapore
City	Singapore
Period	5/07/15 → 9/07/15
Internet address	https://web.archive.org/web/20150728131139/http://www.pacis2015.org/ https://aisel.aisnet.org/pacis2015/ (Proceedings)

Keywords

Large-scale dataset
Novelty detection
One-class Support Vector Machine

Access to Document

10.1007/978-3-319-18032-8_15

Cite this

Le, T., Phung, D., Nguyen, K., & Venkatesh, S. (2015). Fast One-Class support Vector Machine for novelty detection. In T. Cao, E.-P. Lim, Z.-H. Zhou, T.-B. Ho, D. Cheung, & H. Motoda (Eds.), Advances in Knowledge Discovery and Data Mining: 19th Pacific-Asia Conference, PAKDD 2015 Ho Chi Minh City, Vietnam, May 19–22, 2015 Proceedings, Part II (pp. 189-200). (Lecture Notes in Computer Science ; Vol. 9078). Springer. https://doi.org/10.1007/978-3-319-18032-8_15

Le, Trung ; Phung, Dinh ; Nguyen, Khanh et al. / Fast One-Class support Vector Machine for novelty detection. Advances in Knowledge Discovery and Data Mining: 19th Pacific-Asia Conference, PAKDD 2015 Ho Chi Minh City, Vietnam, May 19–22, 2015 Proceedings, Part II. editor / Tru Cao ; Ee-Peng Lim ; Zhi-Hua Zhou ; Tu-Bao Ho ; David Cheung ; Hiroshi Motoda. Cham Switzerland : Springer, 2015. pp. 189-200 (Lecture Notes in Computer Science ).

@inproceedings{0196c3bfc3b74479a303e7e4918a099c,

title = "Fast One-Class support Vector Machine for novelty detection",

abstract = "Novelty detection arises as an important learning task in several applications. Kernel-based approach to novelty detection has been widely used due to its theoretical rigor and elegance of geometric interpretation. However, computational complexity is a major obstacle in this approach. In this paper, leveraging on the cutting-plane framework with the well-known One-Class Support Vector Machine, we present a new solution that can scale up seamlessly with data. The first solution is exact and linear when viewed through the cutting-plane; the second employed a sampling strategy that remarkably has a constant computational complexity defined relatively to the probability of approximation accuracy. Several datasets are benchmarked to demonstrate the credibility of our framework.",

keywords = "Large-scale dataset, Novelty detection, One-class Support Vector Machine",

author = "Trung Le and Dinh Phung and Khanh Nguyen and Svetha Venkatesh",

year = "2015",

doi = "10.1007/978-3-319-18032-8\_15",

language = "English",

isbn = "9783319180311",

series = "Lecture Notes in Computer Science ",

publisher = "Springer",

pages = "189--200",

editor = "Tru Cao and Ee-Peng Lim and Zhi-Hua Zhou and Tu-Bao Ho and David Cheung and Hiroshi Motoda",

booktitle = "Advances in Knowledge Discovery and Data Mining",

address = "Switzerland",

note = "Pacific Asia Conference on Information Systems 2015, PACIS 2015 ; Conference date: 05-07-2015 Through 09-07-2015",

url = "https://web.archive.org/web/20150728131139/http://www.pacis2015.org/, https://aisel.aisnet.org/pacis2015/",

}

Le, T , Phung, D, Nguyen, K & Venkatesh, S 2015, Fast One-Class support Vector Machine for novelty detection. in T Cao, E-P Lim, Z-H Zhou, T-B Ho, D Cheung & H Motoda (eds), Advances in Knowledge Discovery and Data Mining: 19th Pacific-Asia Conference, PAKDD 2015 Ho Chi Minh City, Vietnam, May 19–22, 2015 Proceedings, Part II. Lecture Notes in Computer Science , vol. 9078, Springer, Cham Switzerland, pp. 189-200, Pacific Asia Conference on Information Systems 2015, Singapore, Singapore, 5/07/15. https://doi.org/10.1007/978-3-319-18032-8_15

Fast One-Class support Vector Machine for novelty detection. / Le, Trung ; Phung, Dinh; Nguyen, Khanh et al.
Advances in Knowledge Discovery and Data Mining: 19th Pacific-Asia Conference, PAKDD 2015 Ho Chi Minh City, Vietnam, May 19–22, 2015 Proceedings, Part II. ed. / Tru Cao; Ee-Peng Lim; Zhi-Hua Zhou; Tu-Bao Ho; David Cheung; Hiroshi Motoda. Cham Switzerland: Springer, 2015. p. 189-200 (Lecture Notes in Computer Science ; Vol. 9078).

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

TY - GEN

T1 - Fast One-Class support Vector Machine for novelty detection

AU - Le, Trung

AU - Phung, Dinh

AU - Nguyen, Khanh

AU - Venkatesh, Svetha

N1 - Conference code: 19th

PY - 2015

Y1 - 2015

N2 - Novelty detection arises as an important learning task in several applications. Kernel-based approach to novelty detection has been widely used due to its theoretical rigor and elegance of geometric interpretation. However, computational complexity is a major obstacle in this approach. In this paper, leveraging on the cutting-plane framework with the well-known One-Class Support Vector Machine, we present a new solution that can scale up seamlessly with data. The first solution is exact and linear when viewed through the cutting-plane; the second employed a sampling strategy that remarkably has a constant computational complexity defined relatively to the probability of approximation accuracy. Several datasets are benchmarked to demonstrate the credibility of our framework.

AB - Novelty detection arises as an important learning task in several applications. Kernel-based approach to novelty detection has been widely used due to its theoretical rigor and elegance of geometric interpretation. However, computational complexity is a major obstacle in this approach. In this paper, leveraging on the cutting-plane framework with the well-known One-Class Support Vector Machine, we present a new solution that can scale up seamlessly with data. The first solution is exact and linear when viewed through the cutting-plane; the second employed a sampling strategy that remarkably has a constant computational complexity defined relatively to the probability of approximation accuracy. Several datasets are benchmarked to demonstrate the credibility of our framework.

KW - Large-scale dataset

KW - Novelty detection

KW - One-class Support Vector Machine

UR - https://www.scopus.com/pages/publications/84945552187

U2 - 10.1007/978-3-319-18032-8_15

DO - 10.1007/978-3-319-18032-8_15

M3 - Conference Paper

AN - SCOPUS:84945552187

SN - 9783319180311

T3 - Lecture Notes in Computer Science

SP - 189

EP - 200

BT - Advances in Knowledge Discovery and Data Mining

A2 - Cao, Tru

A2 - Lim, Ee-Peng

A2 - Zhou, Zhi-Hua

A2 - Ho, Tu-Bao

A2 - Cheung, David

A2 - Motoda, Hiroshi

PB - Springer

CY - Cham Switzerland

T2 - Pacific Asia Conference on Information Systems 2015

Y2 - 5 July 2015 through 9 July 2015

ER -

Le T , Phung D, Nguyen K, Venkatesh S. Fast One-Class support Vector Machine for novelty detection. In Cao T, Lim EP, Zhou ZH, Ho TB, Cheung D, Motoda H, editors, Advances in Knowledge Discovery and Data Mining: 19th Pacific-Asia Conference, PAKDD 2015 Ho Chi Minh City, Vietnam, May 19–22, 2015 Proceedings, Part II. Cham Switzerland: Springer. 2015. p. 189-200. (Lecture Notes in Computer Science ). doi: 10.1007/978-3-319-18032-8_15

Fast One-Class support Vector Machine for novelty detection

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Cite this