Generalization of quadratic regularized and standard fuzzy c-means clustering with respect to regularization of hard c-means

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

5 Citations (Scopus)

Abstract

In this paper, the quadratic regularized and standard fuzzy c-means clustering algorithms (qFCM and sFCM) are generalized with respect to hard c-means (HCM) regularization. First, qFCM is generalized from quadratic regularization to power regularization. The relation between this generalization and sFCM is then compared to the relation between other pairs of methods from the perspective of HCM regularization, and, based on this comparison, sFCM is generalized through the addition of a fuzzification parameter. In this process, we see that other methods can be constructed by combining HCM and a regularization term that can either be weighted by data-cluster dissimilarity or not. Furthermore, we see numerically that the existence or nonexistence of this weighting determines the property of these methods' classification rules for an extremely large datum. We also note that the problem of non-convergence in some methods can be avoided through further modification.

Original languageEnglish
Title of host publicationModeling Decisions for Artificial Intelligence - 10th International Conference, MDAI 2013, Proceedings
Pages152-165
Number of pages14
DOIs
Publication statusPublished - 2013 Dec 1
Event10th International Conference on Modeling Decisions for Artificial Intelligence, MDAI 2013 - Barcelona, Spain
Duration: 2013 Nov 202013 Nov 22

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8234 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th International Conference on Modeling Decisions for Artificial Intelligence, MDAI 2013
CountrySpain
CityBarcelona
Period13/11/2013/11/22

    Fingerprint

Keywords

  • fuzzy c-means clustering
  • regularization

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Kanzawa, Y. (2013). Generalization of quadratic regularized and standard fuzzy c-means clustering with respect to regularization of hard c-means. In Modeling Decisions for Artificial Intelligence - 10th International Conference, MDAI 2013, Proceedings (pp. 152-165). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8234 LNAI). https://doi.org/10.1007/978-3-642-41550-0_14