Fuzzy Multi-Sphere Support Vector Data Description

Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013)

Trung Minh LE, Dat TRAN, Wanli MA

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

5 Citations (Scopus)

Abstract

Current well-known data description methods such as Support Vector Data Description and Small Sphere Large Margin are 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 Fuzzy Multi-sphere Support Vector Data Description 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 languageEnglish
Title of host publicationPacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013)
EditorsJian Pei, Vincent S Tseny, Longbird Cao, Hiroshi Motoda, Guandong Xu
Place of PublicationBerlin
PublisherSpringer
Pages570-581
Number of pages12
Volume7819
ISBN (Print)9783642374524
DOIs
Publication statusPublished - 2013
Event17th Pacific-Asia Conference on Knowledge Discovery and Data Mining - Gold Coast, Gold Coast, Australia
Duration: 14 Apr 201317 Apr 2013

Conference

Conference17th Pacific-Asia Conference on Knowledge Discovery and Data Mining
CountryAustralia
CityGold Coast
Period14/04/1317/04/13

Fingerprint

Data description
Data mining

Cite this

LE, T. M., TRAN, D., & MA, W. (2013). Fuzzy Multi-Sphere Support Vector Data Description: Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). In J. Pei, V. S. Tseny, L. Cao, H. Motoda, & G. Xu (Eds.), Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013) (Vol. 7819, pp. 570-581). Berlin: Springer. https://doi.org/10.1007/978-3-642-37456-2_48
LE, Trung Minh ; TRAN, Dat ; MA, Wanli. / Fuzzy Multi-Sphere Support Vector Data Description : Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). editor / Jian Pei ; Vincent S Tseny ; Longbird Cao ; Hiroshi Motoda ; Guandong Xu. Vol. 7819 Berlin : Springer, 2013. pp. 570-581
@inproceedings{444ad92fc9ab455aac69e284db40c076,
title = "Fuzzy Multi-Sphere Support Vector Data Description: Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013)",
abstract = "Current well-known data description methods such as Support Vector Data Description and Small Sphere Large Margin are 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 Fuzzy Multi-sphere Support Vector Data Description 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 = "Support Vector Data Description, Kernel Method, EEG",
author = "LE, {Trung Minh} and Dat TRAN and Wanli MA",
year = "2013",
doi = "10.1007/978-3-642-37456-2_48",
language = "English",
isbn = "9783642374524",
volume = "7819",
pages = "570--581",
editor = "Jian Pei and Tseny, {Vincent S} and Longbird Cao and Hiroshi Motoda and Guandong Xu",
booktitle = "Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013)",
publisher = "Springer",
address = "Netherlands",

}

LE, TM, TRAN, D & MA, W 2013, Fuzzy Multi-Sphere Support Vector Data Description: Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). in J Pei, VS Tseny, L Cao, H Motoda & G Xu (eds), Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). vol. 7819, Springer, Berlin, pp. 570-581, 17th Pacific-Asia Conference on Knowledge Discovery and Data Mining, Gold Coast, Australia, 14/04/13. https://doi.org/10.1007/978-3-642-37456-2_48

Fuzzy Multi-Sphere Support Vector Data Description : Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). / LE, Trung Minh; TRAN, Dat; MA, Wanli.

Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). ed. / Jian Pei; Vincent S Tseny; Longbird Cao; Hiroshi Motoda; Guandong Xu. Vol. 7819 Berlin : Springer, 2013. p. 570-581.

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

TY - GEN

T1 - Fuzzy Multi-Sphere Support Vector Data Description

T2 - Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013)

AU - LE, Trung Minh

AU - TRAN, Dat

AU - MA, Wanli

PY - 2013

Y1 - 2013

N2 - Current well-known data description methods such as Support Vector Data Description and Small Sphere Large Margin are 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 Fuzzy Multi-sphere Support Vector Data Description 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 methods such as Support Vector Data Description and Small Sphere Large Margin are 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 Fuzzy Multi-sphere Support Vector Data Description 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 - Support Vector Data Description

KW - Kernel Method

KW - EEG

U2 - 10.1007/978-3-642-37456-2_48

DO - 10.1007/978-3-642-37456-2_48

M3 - Conference contribution

SN - 9783642374524

VL - 7819

SP - 570

EP - 581

BT - Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013)

A2 - Pei, Jian

A2 - Tseny, Vincent S

A2 - Cao, Longbird

A2 - Motoda, Hiroshi

A2 - Xu, Guandong

PB - Springer

CY - Berlin

ER -

LE TM, TRAN D, MA W. Fuzzy Multi-Sphere Support Vector Data Description: Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). In Pei J, Tseny VS, Cao L, Motoda H, Xu G, editors, Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD 2013). Vol. 7819. Berlin: Springer. 2013. p. 570-581 https://doi.org/10.1007/978-3-642-37456-2_48