Deterministic annealing multi-sphere support vector data description

Trung Le; Dat Tran; Wanli Ma; Dharmendra Sharma

doi:10.1007/978-3-642-34487-9_23

Deterministic annealing multi-sphere support vector data description

Trung Le, Dat Tran, Wanli Ma, Dharmendra Sharma

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

1 Citation (Scopus)

Abstract

Current well-known data description method such as Support Vector Data Description is conducted with assumption that data samples of a class in feature space are drawn from a single distribution. Based on this assumption, a single hypersphere is constructed to provide a good data description for the data. However, real-world data samples may be drawn from some distinctive distributions and hence it does not guarantee that a single hypersphere can offer the best data description. In this paper, we introduce a Deterministic Annealing Multi-sphere Support Vector Data Description (DAMS-SVDD) approach to address this issue. We propose to use a set of hyperspheres to provide a better data description for a given data set. Calculations for determining optimal hyperspheres and experimental results for applying this proposed approach to classification problems are presented.

Original language	English
Title of host publication	Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings
Pages	183-190
Number of pages	8
Edition	PART 3
DOIs	https://doi.org/10.1007/978-3-642-34487-9_23
Publication status	Published - 2012
Externally published	Yes
Event	International Conference on Neural Information Processing 2012 - Doha, Qatar Duration: 12 Nov 2012 → 15 Nov 2012 Conference number: 19th https://link.springer.com/book/10.1007/978-3-642-34500-5 (Proceedings)

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Number	PART 3
Volume	7665
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	International Conference on Neural Information Processing 2012
Abbreviated title	ICONIP 2012
Country/Territory	Qatar
City	Doha
Period	12/11/12 → 15/11/12
Internet address	https://link.springer.com/book/10.1007/978-3-642-34500-5 (Proceedings)

Keywords

Deterministic Annealing
Kernel Methods
Multi-Sphere Support Vector Data Description
Support Vector Data Description

Access to Document

10.1007/978-3-642-34487-9_23

Cite this

@inproceedings{5d74f35fdba24e599d8ea8407b97eb2c,

title = "Deterministic annealing multi-sphere support vector data description",

abstract = "Current well-known data description method such as Support Vector Data Description is conducted with assumption that data samples of a class in feature space are drawn from a single distribution. Based on this assumption, a single hypersphere is constructed to provide a good data description for the data. However, real-world data samples may be drawn from some distinctive distributions and hence it does not guarantee that a single hypersphere can offer the best data description. In this paper, we introduce a Deterministic Annealing Multi-sphere Support Vector Data Description (DAMS-SVDD) approach to address this issue. We propose to use a set of hyperspheres to provide a better data description for a given data set. Calculations for determining optimal hyperspheres and experimental results for applying this proposed approach to classification problems are presented.",

keywords = "Deterministic Annealing, Kernel Methods, Multi-Sphere Support Vector Data Description, Support Vector Data Description",

author = "Trung Le and Dat Tran and Wanli Ma and Dharmendra Sharma",

year = "2012",

doi = "10.1007/978-3-642-34487-9\_23",

language = "English",

isbn = "9783642344862",

series = "Lecture Notes in Computer Science ",

publisher = "Springer",

number = "PART 3",

pages = "183--190",

booktitle = "Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings",

edition = "PART 3",

note = "International Conference on Neural Information Processing 2012, ICONIP 2012 ; Conference date: 12-11-2012 Through 15-11-2012",

url = "https://link.springer.com/book/10.1007/978-3-642-34500-5",

}

Le, T, Tran, D, Ma, W & Sharma, D 2012, Deterministic annealing multi-sphere support vector data description. in Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings. PART 3 edn, Lecture Notes in Computer Science , no. PART 3, vol. 7665 , pp. 183-190, International Conference on Neural Information Processing 2012, Doha, Qatar, 12/11/12. https://doi.org/10.1007/978-3-642-34487-9_23

Deterministic annealing multi-sphere support vector data description. / Le, Trung; Tran, Dat; Ma, Wanli et al.
Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings. PART 3. ed. 2012. p. 183-190 (Lecture Notes in Computer Science ; Vol. 7665 , No. PART 3).

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

TY - GEN

T1 - Deterministic annealing multi-sphere support vector data description

AU - Le, Trung

AU - Tran, Dat

AU - Ma, Wanli

AU - Sharma, Dharmendra

N1 - Conference code: 19th

PY - 2012

Y1 - 2012

N2 - Current well-known data description method such as Support Vector Data Description is conducted with assumption that data samples of a class in feature space are drawn from a single distribution. Based on this assumption, a single hypersphere is constructed to provide a good data description for the data. However, real-world data samples may be drawn from some distinctive distributions and hence it does not guarantee that a single hypersphere can offer the best data description. In this paper, we introduce a Deterministic Annealing Multi-sphere Support Vector Data Description (DAMS-SVDD) approach to address this issue. We propose to use a set of hyperspheres to provide a better data description for a given data set. Calculations for determining optimal hyperspheres and experimental results for applying this proposed approach to classification problems are presented.

AB - Current well-known data description method such as Support Vector Data Description is conducted with assumption that data samples of a class in feature space are drawn from a single distribution. Based on this assumption, a single hypersphere is constructed to provide a good data description for the data. However, real-world data samples may be drawn from some distinctive distributions and hence it does not guarantee that a single hypersphere can offer the best data description. In this paper, we introduce a Deterministic Annealing Multi-sphere Support Vector Data Description (DAMS-SVDD) approach to address this issue. We propose to use a set of hyperspheres to provide a better data description for a given data set. Calculations for determining optimal hyperspheres and experimental results for applying this proposed approach to classification problems are presented.

KW - Deterministic Annealing

KW - Kernel Methods

KW - Multi-Sphere Support Vector Data Description

KW - Support Vector Data Description

UR - https://www.scopus.com/pages/publications/84869061532

U2 - 10.1007/978-3-642-34487-9_23

DO - 10.1007/978-3-642-34487-9_23

M3 - Conference Paper

AN - SCOPUS:84869061532

SN - 9783642344862

T3 - Lecture Notes in Computer Science

SP - 183

EP - 190

BT - Neural Information Processing - 19th International Conference, ICONIP 2012, Proceedings

T2 - International Conference on Neural Information Processing 2012

Y2 - 12 November 2012 through 15 November 2012

ER -

Deterministic annealing multi-sphere support vector data description

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Cite this