A gaussian process robust regression

Noboru Murata, Yusuke Kuroda

    Research output: Contribution to journalArticle

    Abstract

    A modified Gaussian process regression is proposed aiming at making regressors robust against outliers. The proposed method is based on U-loss, which is introduced as a natural extension of Kullback-Leibler divergence. The robustness is examined based on the influence function, and numerical experiments are conducted for contaminated data sets and it is shown that the practical performance agrees with the theoretical analysis.

    Original languageEnglish
    Pages (from-to)280-283
    Number of pages4
    JournalProgress of Theoretical Physics Supplement
    Volume157
    Publication statusPublished - 2005

    Fingerprint

    regression analysis
    divergence

    ASJC Scopus subject areas

    • Physics and Astronomy(all)

    Cite this

    A gaussian process robust regression. / Murata, Noboru; Kuroda, Yusuke.

    In: Progress of Theoretical Physics Supplement, Vol. 157, 2005, p. 280-283.

    Research output: Contribution to journalArticle

    @article{4b907d0a1f5742859a5579161a4b55a8,
    title = "A gaussian process robust regression",
    abstract = "A modified Gaussian process regression is proposed aiming at making regressors robust against outliers. The proposed method is based on U-loss, which is introduced as a natural extension of Kullback-Leibler divergence. The robustness is examined based on the influence function, and numerical experiments are conducted for contaminated data sets and it is shown that the practical performance agrees with the theoretical analysis.",
    author = "Noboru Murata and Yusuke Kuroda",
    year = "2005",
    language = "English",
    volume = "157",
    pages = "280--283",
    journal = "Progress of Theoretical Physics",
    issn = "0033-068X",
    publisher = "Yukawa Institute for Theoretical Physics",

    }

    TY - JOUR

    T1 - A gaussian process robust regression

    AU - Murata, Noboru

    AU - Kuroda, Yusuke

    PY - 2005

    Y1 - 2005

    N2 - A modified Gaussian process regression is proposed aiming at making regressors robust against outliers. The proposed method is based on U-loss, which is introduced as a natural extension of Kullback-Leibler divergence. The robustness is examined based on the influence function, and numerical experiments are conducted for contaminated data sets and it is shown that the practical performance agrees with the theoretical analysis.

    AB - A modified Gaussian process regression is proposed aiming at making regressors robust against outliers. The proposed method is based on U-loss, which is introduced as a natural extension of Kullback-Leibler divergence. The robustness is examined based on the influence function, and numerical experiments are conducted for contaminated data sets and it is shown that the practical performance agrees with the theoretical analysis.

    UR - http://www.scopus.com/inward/record.url?scp=22144464448&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=22144464448&partnerID=8YFLogxK

    M3 - Article

    AN - SCOPUS:22144464448

    VL - 157

    SP - 280

    EP - 283

    JO - Progress of Theoretical Physics

    JF - Progress of Theoretical Physics

    SN - 0033-068X

    ER -