Estimators with Prescribed Precision in Stochastic Regression Models

Victor Konev

    Research output: Contribution to journal › Article

    14 Citations (Scopus)

    Abstract

    This paper presents stopping rules and associated estimators, with prescribed mean squared errors, of the regression parameters in stochastic regression models. The construction makes fundamental use of the martingale structure of least squares estimates or their modifications. For one-dimensional regressors, the stopping rules simply stop as soon as the conditional variance of the underlying martingale exceeds some suitably chosen threshold. We show how this idea can be modified for the case of multidimensional stochastic regressors.

    Original languageEnglish
    Pages (from-to)179-192
    Number of pages14
    JournalSequential Analysis
    Volume14
    Issue number3
    DOIs
    Publication statusPublished - 1 Jan 1995

    Fingerprint

    Stopping Rule
    Martingale
    Stochastic Model
    Regression Model
    Estimator
    Conditional Variance
    Least Squares Estimate
    Mean Squared Error
    Exceed
    Regression

    Keywords

    • autoregression
    • fixed precision estimators
    • martingales
    • sequential estimation
    • stochastic regressors

    ASJC Scopus subject areas

    • Modelling and Simulation
    • Statistics and Probability

    Cite this

    Estimators with Prescribed Precision in Stochastic Regression Models. / Konev, Victor.

    In: Sequential Analysis, Vol. 14, No. 3, 01.01.1995, p. 179-192.

    Research output: Contribution to journal › Article

    Konev, V 1995, 'Estimators with Prescribed Precision in Stochastic Regression Models', Sequential Analysis, vol. 14, no. 3, pp. 179-192. https://doi.org/10.1080/07474949508836330
    Konev V. Estimators with Prescribed Precision in Stochastic Regression Models. Sequential Analysis. 1995 Jan 1;14(3):179-192. https://doi.org/10.1080/07474949508836330
    Konev, Victor. / Estimators with Prescribed Precision in Stochastic Regression Models. In: Sequential Analysis. 1995 ; Vol. 14, No. 3. pp. 179-192.
    @article{f6e6d43ec74f48d19dd339d48494d0b2,
    title = "Estimators with Prescribed Precision in Stochastic Regression Models",
    abstract = "This paper presents stopping rules and associated estimators, with prescribed mean squared errors, of the regression parameters in stochastic regression models. The construction makes fundamental use of the martingale structure of least squares estimates or their modifications. For one-dimensional regressors, the stopping rules simply stop as soon as the conditional variance of the underlying martingale exceeds some suitably chosen threshold. We show how this idea can be modified for the case of multidimensional stochastic regressors.",
    keywords = "autoregression, fixed precision estimators, martingales, sequential estimation, stochastic regressors",
    author = "Victor Konev",
    year = "1995",
    month = "1",
    day = "1",
    doi = "10.1080/07474949508836330",
    language = "English",
    volume = "14",
    pages = "179--192",
    journal = "Sequential Analysis",
    issn = "0747-4946",
    publisher = "Taylor and Francis Ltd.",
    number = "3",

    }

    TY - JOUR

    T1 - Estimators with Prescribed Precision in Stochastic Regression Models

    AU - Konev, Victor

    PY - 1995/1/1

    Y1 - 1995/1/1

    N2 - This paper presents stopping rules and associated estimators, with prescribed mean squared errors, of the regression parameters in stochastic regression models. The construction makes fundamental use of the martingale structure of least squares estimates or their modifications. For one-dimensional regressors, the stopping rules simply stop as soon as the conditional variance of the underlying martingale exceeds some suitably chosen threshold. We show how this idea can be modified for the case of multidimensional stochastic regressors.

    AB - This paper presents stopping rules and associated estimators, with prescribed mean squared errors, of the regression parameters in stochastic regression models. The construction makes fundamental use of the martingale structure of least squares estimates or their modifications. For one-dimensional regressors, the stopping rules simply stop as soon as the conditional variance of the underlying martingale exceeds some suitably chosen threshold. We show how this idea can be modified for the case of multidimensional stochastic regressors.

    KW - autoregression

    KW - fixed precision estimators

    KW - martingales

    KW - sequential estimation

    KW - stochastic regressors

    UR - http://www.scopus.com/inward/record.url?scp=0007255099&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=0007255099&partnerID=8YFLogxK

    U2 - 10.1080/07474949508836330

    DO - 10.1080/07474949508836330

    M3 - Article

    AN - SCOPUS:0007255099

    VL - 14

    SP - 179

    EP - 192

    JO - Sequential Analysis

    JF - Sequential Analysis

    SN - 0747-4946

    IS - 3

    ER -

    Access to Document