A non-parametric maximum entropy clustering

Hideitsu Hino, Noboru Murata

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

    Abstract

    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.

    Original languageEnglish
    Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    PublisherSpringer Verlag
    Pages113-120
    Number of pages8
    Volume8681 LNCS
    ISBN (Print)9783319111780
    DOIs
    Publication statusPublished - 2014
    Event24th International Conference on Artificial Neural Networks, ICANN 2014 - Hamburg, Germany
    Duration: 2014 Sep 152014 Sep 19

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume8681 LNCS
    ISSN (Print)03029743
    ISSN (Electronic)16113349

    Other

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

    Fingerprint

    Maximum Entropy
    Entropy
    Clustering
    Clustering algorithms
    Sampling
    Exploratory Data Analysis
    Nonparametric Methods
    Mutual Information
    Clustering Methods
    Clustering Algorithm
    Optimization

    Keywords

    • Information Theoretic Clustering
    • Likelihood and Entropy estimator
    • Non-parametric

    ASJC Scopus subject areas

    • Computer Science(all)
    • Theoretical Computer Science

    Cite this

    Hino, H., & Murata, N. (2014). A non-parametric maximum entropy clustering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8681 LNCS, pp. 113-120). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8681 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-11179-7_15

    A non-parametric maximum entropy clustering. / Hino, Hideitsu; Murata, Noboru.

    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8681 LNCS Springer Verlag, 2014. p. 113-120 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8681 LNCS).

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

    Hino, H & Murata, N 2014, A non-parametric maximum entropy clustering. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 8681 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8681 LNCS, Springer Verlag, pp. 113-120, 24th International Conference on Artificial Neural Networks, ICANN 2014, Hamburg, Germany, 14/9/15. https://doi.org/10.1007/978-3-319-11179-7_15
    Hino H, Murata N. A non-parametric maximum entropy clustering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8681 LNCS. Springer Verlag. 2014. p. 113-120. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-11179-7_15
    Hino, Hideitsu ; Murata, Noboru. / A non-parametric maximum entropy clustering. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 8681 LNCS Springer Verlag, 2014. pp. 113-120 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
    @inproceedings{c859021a29ca449891136aadc92ded9c,
    title = "A non-parametric maximum entropy clustering",
    abstract = "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.",
    keywords = "Information Theoretic Clustering, Likelihood and Entropy estimator, Non-parametric",
    author = "Hideitsu Hino and Noboru Murata",
    year = "2014",
    doi = "10.1007/978-3-319-11179-7_15",
    language = "English",
    isbn = "9783319111780",
    volume = "8681 LNCS",
    series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
    publisher = "Springer Verlag",
    pages = "113--120",
    booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

    }

    TY - GEN

    T1 - A non-parametric maximum entropy clustering

    AU - Hino, Hideitsu

    AU - Murata, Noboru

    PY - 2014

    Y1 - 2014

    N2 - 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.

    AB - 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.

    KW - Information Theoretic Clustering

    KW - Likelihood and Entropy estimator

    KW - Non-parametric

    UR - http://www.scopus.com/inward/record.url?scp=84958540640&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84958540640&partnerID=8YFLogxK

    U2 - 10.1007/978-3-319-11179-7_15

    DO - 10.1007/978-3-319-11179-7_15

    M3 - Conference contribution

    SN - 9783319111780

    VL - 8681 LNCS

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

    SP - 113

    EP - 120

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

    PB - Springer Verlag

    ER -