Local intrinsic dimension estimation by generalized linear Modeling

Hideitsu Hino, Jun Fujiki, Shotaro Akaho, Noboru Murata

    Research output: Contribution to journalLetter

    2 Citations (Scopus)

    Abstract

    We propose a method for intrinsic dimension estimation. By fitting the power of distance from an inspection point and the number of samples included inside a ball with a radius equal to the distance, to a regression model, we estimate the goodness of fit. Then, by using the maximum likelihood method, we estimate the local intrinsic dimension around the inspection point. The proposed method is shown to be comparable to conventional methods in global intrinsic dimension estimation experiments. Furthermore, we experimentally show that the proposed method outperforms a conventional local dimension estimation method.

    Original languageEnglish
    Pages (from-to)1838-1878
    Number of pages41
    JournalNeural Computation
    Volume29
    Issue number7
    DOIs
    Publication statusPublished - 2017 Jul 1

    Fingerprint

    Likelihood Functions
    Intrinsic
    Modeling
    Conventional
    Goodness
    Maximum Likelihood
    Experiment

    ASJC Scopus subject areas

    • Arts and Humanities (miscellaneous)
    • Cognitive Neuroscience

    Cite this

    Local intrinsic dimension estimation by generalized linear Modeling. / Hino, Hideitsu; Fujiki, Jun; Akaho, Shotaro; Murata, Noboru.

    In: Neural Computation, Vol. 29, No. 7, 01.07.2017, p. 1838-1878.

    Research output: Contribution to journalLetter

    Hino, Hideitsu ; Fujiki, Jun ; Akaho, Shotaro ; Murata, Noboru. / Local intrinsic dimension estimation by generalized linear Modeling. In: Neural Computation. 2017 ; Vol. 29, No. 7. pp. 1838-1878.
    @article{213fadd85c594e5d96efda8eb455ab0c,
    title = "Local intrinsic dimension estimation by generalized linear Modeling",
    abstract = "We propose a method for intrinsic dimension estimation. By fitting the power of distance from an inspection point and the number of samples included inside a ball with a radius equal to the distance, to a regression model, we estimate the goodness of fit. Then, by using the maximum likelihood method, we estimate the local intrinsic dimension around the inspection point. The proposed method is shown to be comparable to conventional methods in global intrinsic dimension estimation experiments. Furthermore, we experimentally show that the proposed method outperforms a conventional local dimension estimation method.",
    author = "Hideitsu Hino and Jun Fujiki and Shotaro Akaho and Noboru Murata",
    year = "2017",
    month = "7",
    day = "1",
    doi = "10.1162/NECO_a_00969",
    language = "English",
    volume = "29",
    pages = "1838--1878",
    journal = "Neural Computation",
    issn = "0899-7667",
    publisher = "MIT Press Journals",
    number = "7",

    }

    TY - JOUR

    T1 - Local intrinsic dimension estimation by generalized linear Modeling

    AU - Hino, Hideitsu

    AU - Fujiki, Jun

    AU - Akaho, Shotaro

    AU - Murata, Noboru

    PY - 2017/7/1

    Y1 - 2017/7/1

    N2 - We propose a method for intrinsic dimension estimation. By fitting the power of distance from an inspection point and the number of samples included inside a ball with a radius equal to the distance, to a regression model, we estimate the goodness of fit. Then, by using the maximum likelihood method, we estimate the local intrinsic dimension around the inspection point. The proposed method is shown to be comparable to conventional methods in global intrinsic dimension estimation experiments. Furthermore, we experimentally show that the proposed method outperforms a conventional local dimension estimation method.

    AB - We propose a method for intrinsic dimension estimation. By fitting the power of distance from an inspection point and the number of samples included inside a ball with a radius equal to the distance, to a regression model, we estimate the goodness of fit. Then, by using the maximum likelihood method, we estimate the local intrinsic dimension around the inspection point. The proposed method is shown to be comparable to conventional methods in global intrinsic dimension estimation experiments. Furthermore, we experimentally show that the proposed method outperforms a conventional local dimension estimation method.

    UR - http://www.scopus.com/inward/record.url?scp=85021381169&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85021381169&partnerID=8YFLogxK

    U2 - 10.1162/NECO_a_00969

    DO - 10.1162/NECO_a_00969

    M3 - Letter

    VL - 29

    SP - 1838

    EP - 1878

    JO - Neural Computation

    JF - Neural Computation

    SN - 0899-7667

    IS - 7

    ER -