Presentation 2014-07-26
Estimating the number of clusters for spherical clustering
Kazuhisa FUJITA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Estimation of the number of clusters is an important issue on data clustering. The one of the major methods to estimate the number of clusters is x-means. The x-means is used for clustering data distributed on Cartesian coordinates. However, x-means cannot be applied to clustering data that is spherically distributed. The purpose of the present study is to restructure x-means to be applied to spherical data. To address the issue, I assume that the data is generated from von Mises-Fisher distribution and restructure x-means. I demonstrate that the proposed method can estimate the number of clustering.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Spherical clustering / Estimation of the number of clusters / k-means / x-means / EM algorithm / von Mises-Fisher distribution
Paper # NC2014-20
Date of Issue

Conference Information
Committee NC
Conference Date 2014/7/19(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Estimating the number of clusters for spherical clustering
Sub Title (in English)
Keyword(1) Spherical clustering
Keyword(2) Estimation of the number of clusters
Keyword(3) k-means
Keyword(4) x-means
Keyword(5) EM algorithm
Keyword(6) von Mises-Fisher distribution
1st Author's Name Kazuhisa FUJITA
1st Author's Affiliation National Institute of Technology, Tsuyama College:University of Electro-Communications()
Date 2014-07-26
Paper # NC2014-20
Volume (vol) vol.114
Number (no) 154
Page pp.pp.-
#Pages 6
Date of Issue