Influence diagnostics in possibly asymmetric circular-linear multivariate regression models

Shuangzhe LIU, Tiefeng Ma, Ashis SenGupta, Kunio Shimizu, Minzhen Wang

    Research output: Contribution to journalArticle

    1 Citation (Scopus)

    Abstract

    Distributional studies and regression models have played important roles in statistical analysis of circular data. Asymmetric circular-linear multivariate regression models (SenGupta and Ugwuowo Environ. Ecol. Stat. 13(3), 299-309 2006) are motivated by and applied to predict some environmental characteristics based on both circular and linear predictors. In this paper, we consider a likelihood approach (Cook J. R. Stat. Soc. Ser. B Stat Methodol. 48(2), 133-169 1986) to study influence diagnostic analysis for these models, using the maximum likelihood estimation and influence diagnostics methods. The observed information matrices and normal curvatures are derived. Simulated and real data examples are then provided to illustrate our approach and establish the utility of our results.

    Original languageEnglish
    Pages (from-to)76-93
    Number of pages18
    JournalSankhya: The Indian Journal of Statistics
    Volume79B
    DOIs
    Publication statusPublished - 2017

    Fingerprint

    Influence Diagnostics
    Multivariate Regression
    Multivariate Models
    Regression Model
    Circular Data
    Normal Curvature
    Observed Information
    Information Matrix
    Maximum Likelihood Estimation
    Statistical Analysis
    Predictors
    Likelihood
    Predict
    Multivariate regression
    Regression model
    Diagnostics
    Model
    Curvature
    Statistical analysis
    Maximum likelihood estimation

    Cite this

    LIU, Shuangzhe ; Ma, Tiefeng ; SenGupta, Ashis ; Shimizu, Kunio ; Wang, Minzhen. / Influence diagnostics in possibly asymmetric circular-linear multivariate regression models. In: Sankhya: The Indian Journal of Statistics. 2017 ; Vol. 79B. pp. 76-93.
    @article{c2e77051296a473d974823322772492a,
    title = "Influence diagnostics in possibly asymmetric circular-linear multivariate regression models",
    abstract = "Distributional studies and regression models have played important roles in statistical analysis of circular data. Asymmetric circular-linear multivariate regression models (SenGupta and Ugwuowo Environ. Ecol. Stat. 13(3), 299-309 2006) are motivated by and applied to predict some environmental characteristics based on both circular and linear predictors. In this paper, we consider a likelihood approach (Cook J. R. Stat. Soc. Ser. B Stat Methodol. 48(2), 133-169 1986) to study influence diagnostic analysis for these models, using the maximum likelihood estimation and influence diagnostics methods. The observed information matrices and normal curvatures are derived. Simulated and real data examples are then provided to illustrate our approach and establish the utility of our results.",
    keywords = "Angular-linear dependency, Maximum likelihood estimation, Solar energy data",
    author = "Shuangzhe LIU and Tiefeng Ma and Ashis SenGupta and Kunio Shimizu and Minzhen Wang",
    year = "2017",
    doi = "10.1007/s13571-016-0116-8",
    language = "English",
    volume = "79B",
    pages = "76--93",
    journal = "Sankhya: The Indian Journal of Statistics",
    issn = "0581-572X",
    publisher = "Indian Statistical Institute",

    }

    Influence diagnostics in possibly asymmetric circular-linear multivariate regression models. / LIU, Shuangzhe; Ma, Tiefeng; SenGupta, Ashis; Shimizu, Kunio; Wang, Minzhen.

    In: Sankhya: The Indian Journal of Statistics, Vol. 79B, 2017, p. 76-93.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - Influence diagnostics in possibly asymmetric circular-linear multivariate regression models

    AU - LIU, Shuangzhe

    AU - Ma, Tiefeng

    AU - SenGupta, Ashis

    AU - Shimizu, Kunio

    AU - Wang, Minzhen

    PY - 2017

    Y1 - 2017

    N2 - Distributional studies and regression models have played important roles in statistical analysis of circular data. Asymmetric circular-linear multivariate regression models (SenGupta and Ugwuowo Environ. Ecol. Stat. 13(3), 299-309 2006) are motivated by and applied to predict some environmental characteristics based on both circular and linear predictors. In this paper, we consider a likelihood approach (Cook J. R. Stat. Soc. Ser. B Stat Methodol. 48(2), 133-169 1986) to study influence diagnostic analysis for these models, using the maximum likelihood estimation and influence diagnostics methods. The observed information matrices and normal curvatures are derived. Simulated and real data examples are then provided to illustrate our approach and establish the utility of our results.

    AB - Distributional studies and regression models have played important roles in statistical analysis of circular data. Asymmetric circular-linear multivariate regression models (SenGupta and Ugwuowo Environ. Ecol. Stat. 13(3), 299-309 2006) are motivated by and applied to predict some environmental characteristics based on both circular and linear predictors. In this paper, we consider a likelihood approach (Cook J. R. Stat. Soc. Ser. B Stat Methodol. 48(2), 133-169 1986) to study influence diagnostic analysis for these models, using the maximum likelihood estimation and influence diagnostics methods. The observed information matrices and normal curvatures are derived. Simulated and real data examples are then provided to illustrate our approach and establish the utility of our results.

    KW - Angular-linear dependency

    KW - Maximum likelihood estimation

    KW - Solar energy data

    UR - http://www.scopus.com/inward/record.url?scp=85022076883&partnerID=8YFLogxK

    U2 - 10.1007/s13571-016-0116-8

    DO - 10.1007/s13571-016-0116-8

    M3 - Article

    VL - 79B

    SP - 76

    EP - 93

    JO - Sankhya: The Indian Journal of Statistics

    JF - Sankhya: The Indian Journal of Statistics

    SN - 0581-572X

    ER -