A lack-of-fit test for generalized linear models via single-index techniques

Chin-Shang Li, Minggen Lu

Research output: Contribution to journalArticle

Abstract

A generalized partially linear single-index model (GPLSIM) is proposed in which the unknown smooth function of single index is approximated by a spline function that can be expressed as a linear combination of B-spline basis functions. The regression coefficients and the unknown smooth function are estimated simultaneously via a modified Fisher-scoring method. It can be shown that the estimators of regression parameters are asymptotically normally distributed. The asymptotic covariance matrix of the estimators can be estimated directly and consistently by using the least-squares method. As an application, the proposed GPLSIM can be employed to assess the lack of fit of a postulated generalized linear model (GLM) based on the comparison of the goodness of fit of the GPLSIM and postulated GLM to construct a likelihood ratio test. An extensive simulation study is conducted to examine the finite-sample performance of the likelihood ratio test. The practicality of the proposed methodology is illustrated with a real-life data set from a study of nesting horseshoe crabs.

Original languageEnglish (US)
Pages (from-to)1-26
Number of pages26
JournalComputational Statistics
DOIs
StateAccepted/In press - Feb 28 2018

Fingerprint

Lack-of-fit Test
Single-index Model
Generalized Linear Model
Linear Model
Likelihood Ratio Test
Smooth function
Fisher Scoring
Lack of Fit
Asymptotic Covariance Matrix
Estimator
B-spline Function
Unknown
Horseshoe
Spline Functions
Splines
Regression Coefficient
Goodness of fit
Least Square Method
Basis Functions
Linear Combination

Keywords

  • B-spline
  • Bootstrap
  • Generalized linear model
  • Generalized partially linear single-index model
  • Likelihood estimator
  • Likelihood ratio test
  • Monte Carlo

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty
  • Computational Mathematics

Cite this

A lack-of-fit test for generalized linear models via single-index techniques. / Li, Chin-Shang; Lu, Minggen.

In: Computational Statistics, 28.02.2018, p. 1-26.

Research output: Contribution to journalArticle

@article{3c655797edb84952a4dcd8846994f539,
title = "A lack-of-fit test for generalized linear models via single-index techniques",
abstract = "A generalized partially linear single-index model (GPLSIM) is proposed in which the unknown smooth function of single index is approximated by a spline function that can be expressed as a linear combination of B-spline basis functions. The regression coefficients and the unknown smooth function are estimated simultaneously via a modified Fisher-scoring method. It can be shown that the estimators of regression parameters are asymptotically normally distributed. The asymptotic covariance matrix of the estimators can be estimated directly and consistently by using the least-squares method. As an application, the proposed GPLSIM can be employed to assess the lack of fit of a postulated generalized linear model (GLM) based on the comparison of the goodness of fit of the GPLSIM and postulated GLM to construct a likelihood ratio test. An extensive simulation study is conducted to examine the finite-sample performance of the likelihood ratio test. The practicality of the proposed methodology is illustrated with a real-life data set from a study of nesting horseshoe crabs.",
keywords = "B-spline, Bootstrap, Generalized linear model, Generalized partially linear single-index model, Likelihood estimator, Likelihood ratio test, Monte Carlo",
author = "Chin-Shang Li and Minggen Lu",
year = "2018",
month = "2",
day = "28",
doi = "10.1007/s00180-018-0802-2",
language = "English (US)",
pages = "1--26",
journal = "Computational Statistics",
issn = "0943-4062",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - A lack-of-fit test for generalized linear models via single-index techniques

AU - Li, Chin-Shang

AU - Lu, Minggen

PY - 2018/2/28

Y1 - 2018/2/28

N2 - A generalized partially linear single-index model (GPLSIM) is proposed in which the unknown smooth function of single index is approximated by a spline function that can be expressed as a linear combination of B-spline basis functions. The regression coefficients and the unknown smooth function are estimated simultaneously via a modified Fisher-scoring method. It can be shown that the estimators of regression parameters are asymptotically normally distributed. The asymptotic covariance matrix of the estimators can be estimated directly and consistently by using the least-squares method. As an application, the proposed GPLSIM can be employed to assess the lack of fit of a postulated generalized linear model (GLM) based on the comparison of the goodness of fit of the GPLSIM and postulated GLM to construct a likelihood ratio test. An extensive simulation study is conducted to examine the finite-sample performance of the likelihood ratio test. The practicality of the proposed methodology is illustrated with a real-life data set from a study of nesting horseshoe crabs.

AB - A generalized partially linear single-index model (GPLSIM) is proposed in which the unknown smooth function of single index is approximated by a spline function that can be expressed as a linear combination of B-spline basis functions. The regression coefficients and the unknown smooth function are estimated simultaneously via a modified Fisher-scoring method. It can be shown that the estimators of regression parameters are asymptotically normally distributed. The asymptotic covariance matrix of the estimators can be estimated directly and consistently by using the least-squares method. As an application, the proposed GPLSIM can be employed to assess the lack of fit of a postulated generalized linear model (GLM) based on the comparison of the goodness of fit of the GPLSIM and postulated GLM to construct a likelihood ratio test. An extensive simulation study is conducted to examine the finite-sample performance of the likelihood ratio test. The practicality of the proposed methodology is illustrated with a real-life data set from a study of nesting horseshoe crabs.

KW - B-spline

KW - Bootstrap

KW - Generalized linear model

KW - Generalized partially linear single-index model

KW - Likelihood estimator

KW - Likelihood ratio test

KW - Monte Carlo

UR - http://www.scopus.com/inward/record.url?scp=85042612558&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042612558&partnerID=8YFLogxK

U2 - 10.1007/s00180-018-0802-2

DO - 10.1007/s00180-018-0802-2

M3 - Article

AN - SCOPUS:85042612558

SP - 1

EP - 26

JO - Computational Statistics

JF - Computational Statistics

SN - 0943-4062

ER -