A non-parametric maximum entropy clustering

Hideitsu Hino, Noboru Murata

研究成果: Conference contribution

抄録

Clustering is a fundamental tool for exploratory data analysis. Information theoretic clustering is based on the optimization of information theoretic quantities such as entropy and mutual information. Recently, since these quantities can be estimated in non-parametric manner, non-parametric information theoretic clustering gains much attention. Assuming the dataset is sampled from a certain cluster, and assigning different sampling weights depending on the clusters, the cluster conditional information theoretic quantities are estimated. In this paper, a simple clustering algorithm is proposed based on the principle of maximum entropy. The algorithm is experimentally shown to be comparable to or outperform conventional non-parametric clustering methods.

本文言語English
ホスト出版物のタイトルArtificial Neural Networks and Machine Learning, ICANN 2014 - 24th International Conference on Artificial Neural Networks, Proceedings
出版社Springer Verlag
ページ113-120
ページ数8
ISBN(印刷版)9783319111780
DOI
出版ステータスPublished - 2014
イベント24th International Conference on Artificial Neural Networks, ICANN 2014 - Hamburg, Germany
継続期間: 2014 9 152014 9 19

出版物シリーズ

名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
8681 LNCS
ISSN(印刷版)0302-9743
ISSN(電子版)1611-3349

Other

Other24th International Conference on Artificial Neural Networks, ICANN 2014
CountryGermany
CityHamburg
Period14/9/1514/9/19

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

フィンガープリント 「A non-parametric maximum entropy clustering」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル