Non-asymptotic tests of model performance

Sylvain Chassang

    Research output: Contribution to journalArticle

    Abstract

    This paper describes a non-asymptotic approach to the problem of selection bias in economic forecasting. By using non-asymptotic measure concentration results, it is possible to deal with settings in which the class of potential models is large with respect to the number of data points. The bounds on p values obtained by these methods are necessarily conservative, but they provide a useful benchmark for model selection in settings where asymptotics may not apply.

    Original languageEnglish (US)
    Pages (from-to)495-514
    Number of pages20
    JournalEconomic Theory
    Volume41
    Issue number3
    DOIs
    StatePublished - Sep 2009

    Fingerprint

    Benchmark
    Economic forecasting
    Model selection
    Selection bias
    P value

    Keywords

    • Model selection
    • Non-asymptotic tests
    • Selection bias

    ASJC Scopus subject areas

    • Economics and Econometrics

    Cite this

    Non-asymptotic tests of model performance. / Chassang, Sylvain.

    In: Economic Theory, Vol. 41, No. 3, 09.2009, p. 495-514.

    Research output: Contribution to journalArticle

    Chassang, Sylvain. / Non-asymptotic tests of model performance. In: Economic Theory. 2009 ; Vol. 41, No. 3. pp. 495-514.
    @article{a337bb749ab24a319856eb551323c430,
    title = "Non-asymptotic tests of model performance",
    abstract = "This paper describes a non-asymptotic approach to the problem of selection bias in economic forecasting. By using non-asymptotic measure concentration results, it is possible to deal with settings in which the class of potential models is large with respect to the number of data points. The bounds on p values obtained by these methods are necessarily conservative, but they provide a useful benchmark for model selection in settings where asymptotics may not apply.",
    keywords = "Model selection, Non-asymptotic tests, Selection bias",
    author = "Sylvain Chassang",
    year = "2009",
    month = "9",
    doi = "10.1007/s00199-008-0408-y",
    language = "English (US)",
    volume = "41",
    pages = "495--514",
    journal = "Economic Theory",
    issn = "0938-2259",
    publisher = "Springer New York",
    number = "3",

    }

    TY - JOUR

    T1 - Non-asymptotic tests of model performance

    AU - Chassang, Sylvain

    PY - 2009/9

    Y1 - 2009/9

    N2 - This paper describes a non-asymptotic approach to the problem of selection bias in economic forecasting. By using non-asymptotic measure concentration results, it is possible to deal with settings in which the class of potential models is large with respect to the number of data points. The bounds on p values obtained by these methods are necessarily conservative, but they provide a useful benchmark for model selection in settings where asymptotics may not apply.

    AB - This paper describes a non-asymptotic approach to the problem of selection bias in economic forecasting. By using non-asymptotic measure concentration results, it is possible to deal with settings in which the class of potential models is large with respect to the number of data points. The bounds on p values obtained by these methods are necessarily conservative, but they provide a useful benchmark for model selection in settings where asymptotics may not apply.

    KW - Model selection

    KW - Non-asymptotic tests

    KW - Selection bias

    UR - http://www.scopus.com/inward/record.url?scp=76349110990&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=76349110990&partnerID=8YFLogxK

    U2 - 10.1007/s00199-008-0408-y

    DO - 10.1007/s00199-008-0408-y

    M3 - Article

    VL - 41

    SP - 495

    EP - 514

    JO - Economic Theory

    JF - Economic Theory

    SN - 0938-2259

    IS - 3

    ER -