On kernelization for a maximizing model of bezdek-like spherical fuzzy c-means clustering

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

7 Citations (Scopus)

Abstract

In this study, we propose three modifications for a maximizing model of spherical Bezdek-type fuzzy c-means clustering (msbFCM). First, we kernelize msbFCM (K-msbFCM). The original msbFCM can only be applied to objects on the first quadrant of the unit hypersphere, whereas its kernelized form can be applied to a wider class of objects. The second modification is a spectral clustering approach to K-msbFCM using a certain assumption. This approach solves the local convergence problem in the original algorithm. The third modification is to construct a model providing the exact solution of the spectral clustering approach. Numerical examples demonstrate that the proposed methods can produce good results for clusters with nonlinear borders when an adequate parameter value is selected.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages108-121
Number of pages14
Volume8825
ISBN (Print)9783319120539
Publication statusPublished - 2014
Event11th International Conference on Modeling Decisions for Artificial Intelligence, MDAI 2014 - Tokyo
Duration: 2014 Oct 292014 Oct 31

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8825
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other11th International Conference on Modeling Decisions for Artificial Intelligence, MDAI 2014
CityTokyo
Period14/10/2914/10/31

Keywords

  • Fuzzy c-means clustering
  • Kernelization
  • Spectral clustering approach

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Kanzawa, Y. (2014). On kernelization for a maximizing model of bezdek-like spherical fuzzy c-means clustering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8825, pp. 108-121). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8825). Springer Verlag.

On kernelization for a maximizing model of bezdek-like spherical fuzzy c-means clustering. / Kanzawa, Yuchi.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8825 Springer Verlag, 2014. p. 108-121 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8825).

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

Kanzawa, Y 2014, On kernelization for a maximizing model of bezdek-like spherical fuzzy c-means clustering. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 8825, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8825, Springer Verlag, pp. 108-121, 11th International Conference on Modeling Decisions for Artificial Intelligence, MDAI 2014, Tokyo , 14/10/29.
Kanzawa Y. On kernelization for a maximizing model of bezdek-like spherical fuzzy c-means clustering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8825. Springer Verlag. 2014. p. 108-121. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Kanzawa, Yuchi. / On kernelization for a maximizing model of bezdek-like spherical fuzzy c-means clustering. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8825 Springer Verlag, 2014. pp. 108-121 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{52821fcdcf904779b5e7287844b383fa,
title = "On kernelization for a maximizing model of bezdek-like spherical fuzzy c-means clustering",
abstract = "In this study, we propose three modifications for a maximizing model of spherical Bezdek-type fuzzy c-means clustering (msbFCM). First, we kernelize msbFCM (K-msbFCM). The original msbFCM can only be applied to objects on the first quadrant of the unit hypersphere, whereas its kernelized form can be applied to a wider class of objects. The second modification is a spectral clustering approach to K-msbFCM using a certain assumption. This approach solves the local convergence problem in the original algorithm. The third modification is to construct a model providing the exact solution of the spectral clustering approach. Numerical examples demonstrate that the proposed methods can produce good results for clusters with nonlinear borders when an adequate parameter value is selected.",
keywords = "Fuzzy c-means clustering, Kernelization, Spectral clustering approach",
author = "Yuchi Kanzawa",
year = "2014",
language = "English",
isbn = "9783319120539",
volume = "8825",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "108--121",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - On kernelization for a maximizing model of bezdek-like spherical fuzzy c-means clustering

AU - Kanzawa, Yuchi

PY - 2014

Y1 - 2014

N2 - In this study, we propose three modifications for a maximizing model of spherical Bezdek-type fuzzy c-means clustering (msbFCM). First, we kernelize msbFCM (K-msbFCM). The original msbFCM can only be applied to objects on the first quadrant of the unit hypersphere, whereas its kernelized form can be applied to a wider class of objects. The second modification is a spectral clustering approach to K-msbFCM using a certain assumption. This approach solves the local convergence problem in the original algorithm. The third modification is to construct a model providing the exact solution of the spectral clustering approach. Numerical examples demonstrate that the proposed methods can produce good results for clusters with nonlinear borders when an adequate parameter value is selected.

AB - In this study, we propose three modifications for a maximizing model of spherical Bezdek-type fuzzy c-means clustering (msbFCM). First, we kernelize msbFCM (K-msbFCM). The original msbFCM can only be applied to objects on the first quadrant of the unit hypersphere, whereas its kernelized form can be applied to a wider class of objects. The second modification is a spectral clustering approach to K-msbFCM using a certain assumption. This approach solves the local convergence problem in the original algorithm. The third modification is to construct a model providing the exact solution of the spectral clustering approach. Numerical examples demonstrate that the proposed methods can produce good results for clusters with nonlinear borders when an adequate parameter value is selected.

KW - Fuzzy c-means clustering

KW - Kernelization

KW - Spectral clustering approach

UR - http://www.scopus.com/inward/record.url?scp=84911046000&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84911046000&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84911046000

SN - 9783319120539

VL - 8825

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

SP - 108

EP - 121

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

PB - Springer Verlag

ER -