Testing lack-of-fit of parametric regression models using nonparametric regression techniques

Randall L. Eubank, Chin-Shang Li, Suojin Wang

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Data-driven lack-of-fit tests are derived for parametric regression models using fit comparison statistics that are based on nonparametric linear smoothers. The tests are applicable to settings where the usual bandwidth/smoothing parameter asymptotics apply to the null model, which includes testing for nonlinear models and some linear models. Large sample distribution theory is established for tests constructed from both kernel and series type estimators. Both types of smoothers are shown to give consistent tests that are asymptotically normal under the null model after appropriate centering and scaling. However, the projection nature of series smoothers results in a simplified scaling factor that produces computational savings for the associated tests.

Original languageEnglish (US)
Pages (from-to)135-152
Number of pages18
JournalStatistica Sinica
Volume15
Issue number1
StatePublished - Jan 2005
Externally publishedYes

Fingerprint

Lack of Fit
Parametric Regression
Nonparametric Regression
Parametric Model
Regression Model
Testing
Null
Lack-of-fit Test
Consistent Test
Large Sample Theory
Distribution Theory
Scaling Factor
Series
Smoothing Parameter
Data-driven
Nonlinear Model
Linear Model
Bandwidth
Projection
Scaling

Keywords

  • Bandwidth selection
  • Fit comparison test
  • Kernel smoother
  • Least squares
  • Series smoother

ASJC Scopus subject areas

  • Mathematics(all)
  • Statistics and Probability

Cite this

Testing lack-of-fit of parametric regression models using nonparametric regression techniques. / Eubank, Randall L.; Li, Chin-Shang; Wang, Suojin.

In: Statistica Sinica, Vol. 15, No. 1, 01.2005, p. 135-152.

Research output: Contribution to journalArticle

@article{9843594f75fa4e2ca9f751848cce2217,
title = "Testing lack-of-fit of parametric regression models using nonparametric regression techniques",
abstract = "Data-driven lack-of-fit tests are derived for parametric regression models using fit comparison statistics that are based on nonparametric linear smoothers. The tests are applicable to settings where the usual bandwidth/smoothing parameter asymptotics apply to the null model, which includes testing for nonlinear models and some linear models. Large sample distribution theory is established for tests constructed from both kernel and series type estimators. Both types of smoothers are shown to give consistent tests that are asymptotically normal under the null model after appropriate centering and scaling. However, the projection nature of series smoothers results in a simplified scaling factor that produces computational savings for the associated tests.",
keywords = "Bandwidth selection, Fit comparison test, Kernel smoother, Least squares, Series smoother",
author = "Eubank, {Randall L.} and Chin-Shang Li and Suojin Wang",
year = "2005",
month = "1",
language = "English (US)",
volume = "15",
pages = "135--152",
journal = "Statistica Sinica",
issn = "1017-0405",
publisher = "Institute of Statistical Science",
number = "1",

}

TY - JOUR

T1 - Testing lack-of-fit of parametric regression models using nonparametric regression techniques

AU - Eubank, Randall L.

AU - Li, Chin-Shang

AU - Wang, Suojin

PY - 2005/1

Y1 - 2005/1

N2 - Data-driven lack-of-fit tests are derived for parametric regression models using fit comparison statistics that are based on nonparametric linear smoothers. The tests are applicable to settings where the usual bandwidth/smoothing parameter asymptotics apply to the null model, which includes testing for nonlinear models and some linear models. Large sample distribution theory is established for tests constructed from both kernel and series type estimators. Both types of smoothers are shown to give consistent tests that are asymptotically normal under the null model after appropriate centering and scaling. However, the projection nature of series smoothers results in a simplified scaling factor that produces computational savings for the associated tests.

AB - Data-driven lack-of-fit tests are derived for parametric regression models using fit comparison statistics that are based on nonparametric linear smoothers. The tests are applicable to settings where the usual bandwidth/smoothing parameter asymptotics apply to the null model, which includes testing for nonlinear models and some linear models. Large sample distribution theory is established for tests constructed from both kernel and series type estimators. Both types of smoothers are shown to give consistent tests that are asymptotically normal under the null model after appropriate centering and scaling. However, the projection nature of series smoothers results in a simplified scaling factor that produces computational savings for the associated tests.

KW - Bandwidth selection

KW - Fit comparison test

KW - Kernel smoother

KW - Least squares

KW - Series smoother

UR - http://www.scopus.com/inward/record.url?scp=17244379745&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=17244379745&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:17244379745

VL - 15

SP - 135

EP - 152

JO - Statistica Sinica

JF - Statistica Sinica

SN - 1017-0405

IS - 1

ER -