An improved Cp criterion for spline smoothing

Chun-Shu Chen, Hsin Cheng Huang

Research output: Contribution to journalArticle

8 Citations (Scopus)

Abstract

Spline smoothing is a popular technique for curve fitting, in which selection of the smoothing parameter is crucial. Many methods such as Mallows' Cp, generalized maximum likelihood (GML), and the extended exponential (EE) criterion have been proposed to select this parameter. Although Cp is shown to be asymptotically optimal, it is usually outperformed by other selection criteria for small to moderate sample sizes due to its high variability. On the other hand, GML and EE are more stable than Cp, but they do not possess the same asymptotic optimality as Cp. Instead of selecting this smoothing parameter directly using Cp, we propose to select among a small class of selection criteria based on Stein's unbiased risk estimate (SURE). Due to the selection effect, the spline estimate obtained from a criterion in this class is nonlinear. Thus, the effective degrees of freedom in SURE contains an adjustment term in addition to the trace of the smoothing matrix, which cannot be ignored in small to moderate sample sizes. The resulting criterion, which we call adaptive Cp, is shown to have an analytic expression, and hence can be efficiently computed. Moreover, adaptive Cp is not only demonstrated to be superior and more stable than commonly used selection criteria in a simulation study, but also shown to possess the same asymptotic optimality as Cp.

Original languageEnglish
Pages (from-to)445-452
Number of pages8
JournalJournal of Statistical Planning and Inference
Volume141
Issue number1
DOIs
Publication statusPublished - 2011 Jan 1

Fingerprint

Spline Smoothing
Splines
Maximum likelihood
Asymptotic Optimality
Smoothing Parameter
Maximum Likelihood
Sample Size
Curve fitting
Estimate
Asymptotically Optimal
Spline
Smoothing
Adjustment
Degree of freedom
Trace
Simulation Study
Term
Selection criteria
Smoothing splines
Class

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Statistics, Probability and Uncertainty
  • Applied Mathematics

Cite this

@article{2f694d5b47e34d98a76226805262a221,
title = "An improved Cp criterion for spline smoothing",
abstract = "Spline smoothing is a popular technique for curve fitting, in which selection of the smoothing parameter is crucial. Many methods such as Mallows' Cp, generalized maximum likelihood (GML), and the extended exponential (EE) criterion have been proposed to select this parameter. Although Cp is shown to be asymptotically optimal, it is usually outperformed by other selection criteria for small to moderate sample sizes due to its high variability. On the other hand, GML and EE are more stable than Cp, but they do not possess the same asymptotic optimality as Cp. Instead of selecting this smoothing parameter directly using Cp, we propose to select among a small class of selection criteria based on Stein's unbiased risk estimate (SURE). Due to the selection effect, the spline estimate obtained from a criterion in this class is nonlinear. Thus, the effective degrees of freedom in SURE contains an adjustment term in addition to the trace of the smoothing matrix, which cannot be ignored in small to moderate sample sizes. The resulting criterion, which we call adaptive Cp, is shown to have an analytic expression, and hence can be efficiently computed. Moreover, adaptive Cp is not only demonstrated to be superior and more stable than commonly used selection criteria in a simulation study, but also shown to possess the same asymptotic optimality as Cp.",
author = "Chun-Shu Chen and Huang, {Hsin Cheng}",
year = "2011",
month = "1",
day = "1",
doi = "10.1016/j.jspi.2010.06.017",
language = "English",
volume = "141",
pages = "445--452",
journal = "Journal of Statistical Planning and Inference",
issn = "0378-3758",
publisher = "Elsevier",
number = "1",

}

An improved Cp criterion for spline smoothing. / Chen, Chun-Shu; Huang, Hsin Cheng.

In: Journal of Statistical Planning and Inference, Vol. 141, No. 1, 01.01.2011, p. 445-452.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An improved Cp criterion for spline smoothing

AU - Chen, Chun-Shu

AU - Huang, Hsin Cheng

PY - 2011/1/1

Y1 - 2011/1/1

N2 - Spline smoothing is a popular technique for curve fitting, in which selection of the smoothing parameter is crucial. Many methods such as Mallows' Cp, generalized maximum likelihood (GML), and the extended exponential (EE) criterion have been proposed to select this parameter. Although Cp is shown to be asymptotically optimal, it is usually outperformed by other selection criteria for small to moderate sample sizes due to its high variability. On the other hand, GML and EE are more stable than Cp, but they do not possess the same asymptotic optimality as Cp. Instead of selecting this smoothing parameter directly using Cp, we propose to select among a small class of selection criteria based on Stein's unbiased risk estimate (SURE). Due to the selection effect, the spline estimate obtained from a criterion in this class is nonlinear. Thus, the effective degrees of freedom in SURE contains an adjustment term in addition to the trace of the smoothing matrix, which cannot be ignored in small to moderate sample sizes. The resulting criterion, which we call adaptive Cp, is shown to have an analytic expression, and hence can be efficiently computed. Moreover, adaptive Cp is not only demonstrated to be superior and more stable than commonly used selection criteria in a simulation study, but also shown to possess the same asymptotic optimality as Cp.

AB - Spline smoothing is a popular technique for curve fitting, in which selection of the smoothing parameter is crucial. Many methods such as Mallows' Cp, generalized maximum likelihood (GML), and the extended exponential (EE) criterion have been proposed to select this parameter. Although Cp is shown to be asymptotically optimal, it is usually outperformed by other selection criteria for small to moderate sample sizes due to its high variability. On the other hand, GML and EE are more stable than Cp, but they do not possess the same asymptotic optimality as Cp. Instead of selecting this smoothing parameter directly using Cp, we propose to select among a small class of selection criteria based on Stein's unbiased risk estimate (SURE). Due to the selection effect, the spline estimate obtained from a criterion in this class is nonlinear. Thus, the effective degrees of freedom in SURE contains an adjustment term in addition to the trace of the smoothing matrix, which cannot be ignored in small to moderate sample sizes. The resulting criterion, which we call adaptive Cp, is shown to have an analytic expression, and hence can be efficiently computed. Moreover, adaptive Cp is not only demonstrated to be superior and more stable than commonly used selection criteria in a simulation study, but also shown to possess the same asymptotic optimality as Cp.

UR - http://www.scopus.com/inward/record.url?scp=77956281310&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77956281310&partnerID=8YFLogxK

U2 - 10.1016/j.jspi.2010.06.017

DO - 10.1016/j.jspi.2010.06.017

M3 - Article

VL - 141

SP - 445

EP - 452

JO - Journal of Statistical Planning and Inference

JF - Journal of Statistical Planning and Inference

SN - 0378-3758

IS - 1

ER -