Testing for no effect via splines

Research output: Contribution to journalArticle

Abstract

A spline-based test statistic for a constant mean function is proposed based on the penalized residual sum-of-squares difference between the null model and a B-spline model in which the regression function is approximated with P-splines approach. When the number of knots is fixed, the limiting null distribution of the test statistic is shown to be the distribution of a linear combination of independent chi-squared random variables, each with one degree of freedom. A smoothing parameter is selected by setting a specified value equal to the expected value of the test statistic under the null hypothesis. Simulation experiments are conducted to study the proposed spline-based test statistic's finite-sample properties.

Original languageEnglish (US)
Pages (from-to)343-357
Number of pages15
JournalComputational Statistics
Volume27
Issue number2
DOIs
StatePublished - Jun 2012

Fingerprint

Splines
Spline
Test Statistic
Statistics
Testing
P-splines
Chi-squared
Smoothing Parameter
Null Distribution
Sum of squares
Regression Function
B-spline
Limiting Distribution
Expected Value
Random variables
Null hypothesis
Knot
Simulation Experiment
Null
Linear Combination

Keywords

  • B-spliness
  • P-splines
  • Penalized least-squares
  • Spline-based test

ASJC Scopus subject areas

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

Cite this

Testing for no effect via splines. / Li, Chin-Shang.

In: Computational Statistics, Vol. 27, No. 2, 06.2012, p. 343-357.

Research output: Contribution to journalArticle

@article{023de2b72b924a6fb25454f4066843a3,
title = "Testing for no effect via splines",
abstract = "A spline-based test statistic for a constant mean function is proposed based on the penalized residual sum-of-squares difference between the null model and a B-spline model in which the regression function is approximated with P-splines approach. When the number of knots is fixed, the limiting null distribution of the test statistic is shown to be the distribution of a linear combination of independent chi-squared random variables, each with one degree of freedom. A smoothing parameter is selected by setting a specified value equal to the expected value of the test statistic under the null hypothesis. Simulation experiments are conducted to study the proposed spline-based test statistic's finite-sample properties.",
keywords = "B-spliness, P-splines, Penalized least-squares, Spline-based test",
author = "Chin-Shang Li",
year = "2012",
month = "6",
doi = "10.1007/s00180-011-0260-6",
language = "English (US)",
volume = "27",
pages = "343--357",
journal = "Computational Statistics",
issn = "0943-4062",
publisher = "Springer Verlag",
number = "2",

}

TY - JOUR

T1 - Testing for no effect via splines

AU - Li, Chin-Shang

PY - 2012/6

Y1 - 2012/6

N2 - A spline-based test statistic for a constant mean function is proposed based on the penalized residual sum-of-squares difference between the null model and a B-spline model in which the regression function is approximated with P-splines approach. When the number of knots is fixed, the limiting null distribution of the test statistic is shown to be the distribution of a linear combination of independent chi-squared random variables, each with one degree of freedom. A smoothing parameter is selected by setting a specified value equal to the expected value of the test statistic under the null hypothesis. Simulation experiments are conducted to study the proposed spline-based test statistic's finite-sample properties.

AB - A spline-based test statistic for a constant mean function is proposed based on the penalized residual sum-of-squares difference between the null model and a B-spline model in which the regression function is approximated with P-splines approach. When the number of knots is fixed, the limiting null distribution of the test statistic is shown to be the distribution of a linear combination of independent chi-squared random variables, each with one degree of freedom. A smoothing parameter is selected by setting a specified value equal to the expected value of the test statistic under the null hypothesis. Simulation experiments are conducted to study the proposed spline-based test statistic's finite-sample properties.

KW - B-spliness

KW - P-splines

KW - Penalized least-squares

KW - Spline-based test

UR - http://www.scopus.com/inward/record.url?scp=84860360538&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84860360538&partnerID=8YFLogxK

U2 - 10.1007/s00180-011-0260-6

DO - 10.1007/s00180-011-0260-6

M3 - Article

VL - 27

SP - 343

EP - 357

JO - Computational Statistics

JF - Computational Statistics

SN - 0943-4062

IS - 2

ER -