Highly sparse reductions to kernel spectral clustering

RaghvenPhDa Mall, Rocco Langone, Johan A.K. Suykens

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Kernel spectral clustering is a model-based spectral clustering method formulated in a primal-dual framework. It has a powerful out-of-sample extension property and a model selection procedure based on the balanced line fit criterion. This paper is an improvement of a previous work which sparsified the kernel spectral clustering method using the line structure of the data projections in the eigenspace. However, the previous method works only in the case of well formed and well separated clusters as in other cases the line structure is lost. In this paper, we propose two highly sparse extensions of kernel spectral clustering that can overcome these limitations. For the selection of the reduced set we use the concept of angles between the data projections in the eigenspace. We show the effectiveness and the amount of sparsity obtained by the proposed methods for several synthetic and real world datasets.

Original languageEnglish
Title of host publicationPattern Recognition and Machine Intelligence - 5th International Conference, PReMI 2013, Proceedings
Pages163-169
Number of pages7
DOIs
Publication statusPublished - 1 Dec 2013
Externally publishedYes
Event5th International Conference on Pattern Recognition and Machine Intelligence, PReMI 2013 - Kolkata, India
Duration: 10 Dec 201314 Dec 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8251 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other5th International Conference on Pattern Recognition and Machine Intelligence, PReMI 2013
CountryIndia
CityKolkata
Period10/12/1314/12/13

Fingerprint

Spectral Clustering
kernel
Eigenspace
Spectral Methods
Clustering Methods
Line
Projection
Model-based Clustering
Primal-dual
Selection Procedures
Sparsity
Model Selection
Angle

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Mall, R., Langone, R., & Suykens, J. A. K. (2013). Highly sparse reductions to kernel spectral clustering. In Pattern Recognition and Machine Intelligence - 5th International Conference, PReMI 2013, Proceedings (pp. 163-169). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8251 LNCS). https://doi.org/10.1007/978-3-642-45062-4_22

Highly sparse reductions to kernel spectral clustering. / Mall, RaghvenPhDa; Langone, Rocco; Suykens, Johan A.K.

Pattern Recognition and Machine Intelligence - 5th International Conference, PReMI 2013, Proceedings. 2013. p. 163-169 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8251 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Mall, R, Langone, R & Suykens, JAK 2013, Highly sparse reductions to kernel spectral clustering. in Pattern Recognition and Machine Intelligence - 5th International Conference, PReMI 2013, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8251 LNCS, pp. 163-169, 5th International Conference on Pattern Recognition and Machine Intelligence, PReMI 2013, Kolkata, India, 10/12/13. https://doi.org/10.1007/978-3-642-45062-4_22
Mall R, Langone R, Suykens JAK. Highly sparse reductions to kernel spectral clustering. In Pattern Recognition and Machine Intelligence - 5th International Conference, PReMI 2013, Proceedings. 2013. p. 163-169. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-45062-4_22
Mall, RaghvenPhDa ; Langone, Rocco ; Suykens, Johan A.K. / Highly sparse reductions to kernel spectral clustering. Pattern Recognition and Machine Intelligence - 5th International Conference, PReMI 2013, Proceedings. 2013. pp. 163-169 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{fd899c4fcf6645daa7d94f7377a55cd7,
title = "Highly sparse reductions to kernel spectral clustering",
abstract = "Kernel spectral clustering is a model-based spectral clustering method formulated in a primal-dual framework. It has a powerful out-of-sample extension property and a model selection procedure based on the balanced line fit criterion. This paper is an improvement of a previous work which sparsified the kernel spectral clustering method using the line structure of the data projections in the eigenspace. However, the previous method works only in the case of well formed and well separated clusters as in other cases the line structure is lost. In this paper, we propose two highly sparse extensions of kernel spectral clustering that can overcome these limitations. For the selection of the reduced set we use the concept of angles between the data projections in the eigenspace. We show the effectiveness and the amount of sparsity obtained by the proposed methods for several synthetic and real world datasets.",
author = "RaghvenPhDa Mall and Rocco Langone and Suykens, {Johan A.K.}",
year = "2013",
month = "12",
day = "1",
doi = "10.1007/978-3-642-45062-4_22",
language = "English",
isbn = "9783642450617",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "163--169",
booktitle = "Pattern Recognition and Machine Intelligence - 5th International Conference, PReMI 2013, Proceedings",

}

TY - GEN

T1 - Highly sparse reductions to kernel spectral clustering

AU - Mall, RaghvenPhDa

AU - Langone, Rocco

AU - Suykens, Johan A.K.

PY - 2013/12/1

Y1 - 2013/12/1

N2 - Kernel spectral clustering is a model-based spectral clustering method formulated in a primal-dual framework. It has a powerful out-of-sample extension property and a model selection procedure based on the balanced line fit criterion. This paper is an improvement of a previous work which sparsified the kernel spectral clustering method using the line structure of the data projections in the eigenspace. However, the previous method works only in the case of well formed and well separated clusters as in other cases the line structure is lost. In this paper, we propose two highly sparse extensions of kernel spectral clustering that can overcome these limitations. For the selection of the reduced set we use the concept of angles between the data projections in the eigenspace. We show the effectiveness and the amount of sparsity obtained by the proposed methods for several synthetic and real world datasets.

AB - Kernel spectral clustering is a model-based spectral clustering method formulated in a primal-dual framework. It has a powerful out-of-sample extension property and a model selection procedure based on the balanced line fit criterion. This paper is an improvement of a previous work which sparsified the kernel spectral clustering method using the line structure of the data projections in the eigenspace. However, the previous method works only in the case of well formed and well separated clusters as in other cases the line structure is lost. In this paper, we propose two highly sparse extensions of kernel spectral clustering that can overcome these limitations. For the selection of the reduced set we use the concept of angles between the data projections in the eigenspace. We show the effectiveness and the amount of sparsity obtained by the proposed methods for several synthetic and real world datasets.

UR - http://www.scopus.com/inward/record.url?scp=84893382878&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84893382878&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-45062-4_22

DO - 10.1007/978-3-642-45062-4_22

M3 - Conference contribution

SN - 9783642450617

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

SP - 163

EP - 169

BT - Pattern Recognition and Machine Intelligence - 5th International Conference, PReMI 2013, Proceedings

ER -