Nonparametric F-tests for nested global and local polynomial models

Li Shan Huang, Haiyan Su

Research output: Contribution to journalArticle

Abstract

In this paper, we investigate geometric properties of local polynomial regression and show that the class of global polynomial models is nested within the class of functions generated by fitting local polynomials. The geometric properties are then utilized to construct nonparametric F-tests for testing whether a regression relationship is a polynomial function. The proposed F-tests can be seen as a "calculus" extension of the classical F-tests with analysis of variance interpretations. With the normality assumption, the test statistic is shown to have asymptotic F-distributions under the null hypothesis and fixed alternatives. Simulation results illustrate that the asymptotic null F-distribution approximates well in finite sample cases and the proposed tests enjoy robustness against heteroscedasticity and non-normality as do the classical F-tests.

Original languageEnglish
Pages (from-to)1372-1380
Number of pages9
JournalJournal of Statistical Planning and Inference
Volume139
Issue number4
DOIs
StatePublished - 1 Apr 2009

Fingerprint

Local Polynomial
F Test
Polynomial Model
Polynomials
F distribution
Local Polynomial Fitting
Local Polynomial Regression
Analysis of variance (ANOVA)
Non-normality
Heteroscedasticity
Null Distribution
Analysis of variance
Polynomial function
Null hypothesis
Statistics
Normality
Asymptotic distribution
Test Statistic
Calculus
Regression

Keywords

  • Local polynomial regression
  • Model checking
  • Nonparametric regression
  • Pseudolikelihood test

Cite this

@article{37a05039c6e745908fbcf538d7f63265,
title = "Nonparametric F-tests for nested global and local polynomial models",
abstract = "In this paper, we investigate geometric properties of local polynomial regression and show that the class of global polynomial models is nested within the class of functions generated by fitting local polynomials. The geometric properties are then utilized to construct nonparametric F-tests for testing whether a regression relationship is a polynomial function. The proposed F-tests can be seen as a {"}calculus{"} extension of the classical F-tests with analysis of variance interpretations. With the normality assumption, the test statistic is shown to have asymptotic F-distributions under the null hypothesis and fixed alternatives. Simulation results illustrate that the asymptotic null F-distribution approximates well in finite sample cases and the proposed tests enjoy robustness against heteroscedasticity and non-normality as do the classical F-tests.",
keywords = "Local polynomial regression, Model checking, Nonparametric regression, Pseudolikelihood test",
author = "Huang, {Li Shan} and Haiyan Su",
year = "2009",
month = "4",
day = "1",
doi = "10.1016/j.jspi.2008.08.013",
language = "English",
volume = "139",
pages = "1372--1380",
journal = "Journal of Statistical Planning and Inference",
issn = "0378-3758",
publisher = "Elsevier",
number = "4",

}

Nonparametric F-tests for nested global and local polynomial models. / Huang, Li Shan; Su, Haiyan.

In: Journal of Statistical Planning and Inference, Vol. 139, No. 4, 01.04.2009, p. 1372-1380.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Nonparametric F-tests for nested global and local polynomial models

AU - Huang, Li Shan

AU - Su, Haiyan

PY - 2009/4/1

Y1 - 2009/4/1

N2 - In this paper, we investigate geometric properties of local polynomial regression and show that the class of global polynomial models is nested within the class of functions generated by fitting local polynomials. The geometric properties are then utilized to construct nonparametric F-tests for testing whether a regression relationship is a polynomial function. The proposed F-tests can be seen as a "calculus" extension of the classical F-tests with analysis of variance interpretations. With the normality assumption, the test statistic is shown to have asymptotic F-distributions under the null hypothesis and fixed alternatives. Simulation results illustrate that the asymptotic null F-distribution approximates well in finite sample cases and the proposed tests enjoy robustness against heteroscedasticity and non-normality as do the classical F-tests.

AB - In this paper, we investigate geometric properties of local polynomial regression and show that the class of global polynomial models is nested within the class of functions generated by fitting local polynomials. The geometric properties are then utilized to construct nonparametric F-tests for testing whether a regression relationship is a polynomial function. The proposed F-tests can be seen as a "calculus" extension of the classical F-tests with analysis of variance interpretations. With the normality assumption, the test statistic is shown to have asymptotic F-distributions under the null hypothesis and fixed alternatives. Simulation results illustrate that the asymptotic null F-distribution approximates well in finite sample cases and the proposed tests enjoy robustness against heteroscedasticity and non-normality as do the classical F-tests.

KW - Local polynomial regression

KW - Model checking

KW - Nonparametric regression

KW - Pseudolikelihood test

UR - http://www.scopus.com/inward/record.url?scp=57749181780&partnerID=8YFLogxK

U2 - 10.1016/j.jspi.2008.08.013

DO - 10.1016/j.jspi.2008.08.013

M3 - Article

AN - SCOPUS:57749181780

VL - 139

SP - 1372

EP - 1380

JO - Journal of Statistical Planning and Inference

JF - Journal of Statistical Planning and Inference

SN - 0378-3758

IS - 4

ER -