Data driven adaptive spline smoothing

Ziyue Liu, Wensheng Guo

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

In classical smoothing splines, the smoothness is controlled by a single smoothing parameter that penalizes the roughness uniformly across the whole domain. Adaptive smoothing splines extend this framework to allow the smoothing parameter to change in the domain, adapting to the change of roughness. In this article we propose a data driven method to nonparametrically model the penalty function. We propose to approximate the penalty function by a step function whose segmentation is data driven, and to estimate it by maximizing the generalized likelihood. A complexity penalty is added to the generalized likelihood in selecting the best step function from a collection of candidates. A state space representation for the adaptive smoothing splines is derived to ease the computational demand. To allow for fast search among the candidate models, we impose a binary tree structure on the penalty function and propose an efficient search algorithm. We show the consistency of the final estimate. We demonstrate the effectiveness of the method through simulations and a data example.

Original languageEnglish (US)
Pages (from-to)1143-1163
Number of pages21
JournalStatistica Sinica
Volume20
Issue number3
StatePublished - Jul 2010
Externally publishedYes

Fingerprint

Adaptive Smoothing
Spline Smoothing
Smoothing Splines
Penalty Function
Data-driven
Step function
Smoothing Parameter
Roughness
Likelihood
State-space Representation
Binary Tree
Tree Structure
Estimate
Search Algorithm
Penalty
Smoothness
Efficient Algorithms
Segmentation
Model
Demonstrate

Keywords

  • Binary tree
  • Complexity penalty
  • Generalized maximum likelihood
  • Model selection
  • State space method

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Liu, Z., & Guo, W. (2010). Data driven adaptive spline smoothing. Statistica Sinica, 20(3), 1143-1163.

Data driven adaptive spline smoothing. / Liu, Ziyue; Guo, Wensheng.

In: Statistica Sinica, Vol. 20, No. 3, 07.2010, p. 1143-1163.

Research output: Contribution to journalArticle

Liu, Z & Guo, W 2010, 'Data driven adaptive spline smoothing', Statistica Sinica, vol. 20, no. 3, pp. 1143-1163.
Liu, Ziyue ; Guo, Wensheng. / Data driven adaptive spline smoothing. In: Statistica Sinica. 2010 ; Vol. 20, No. 3. pp. 1143-1163.
@article{733c1a396970456791955335833b30c9,
title = "Data driven adaptive spline smoothing",
abstract = "In classical smoothing splines, the smoothness is controlled by a single smoothing parameter that penalizes the roughness uniformly across the whole domain. Adaptive smoothing splines extend this framework to allow the smoothing parameter to change in the domain, adapting to the change of roughness. In this article we propose a data driven method to nonparametrically model the penalty function. We propose to approximate the penalty function by a step function whose segmentation is data driven, and to estimate it by maximizing the generalized likelihood. A complexity penalty is added to the generalized likelihood in selecting the best step function from a collection of candidates. A state space representation for the adaptive smoothing splines is derived to ease the computational demand. To allow for fast search among the candidate models, we impose a binary tree structure on the penalty function and propose an efficient search algorithm. We show the consistency of the final estimate. We demonstrate the effectiveness of the method through simulations and a data example.",
keywords = "Binary tree, Complexity penalty, Generalized maximum likelihood, Model selection, State space method",
author = "Ziyue Liu and Wensheng Guo",
year = "2010",
month = "7",
language = "English (US)",
volume = "20",
pages = "1143--1163",
journal = "Statistica Sinica",
issn = "1017-0405",
publisher = "Institute of Statistical Science",
number = "3",

}

TY - JOUR

T1 - Data driven adaptive spline smoothing

AU - Liu, Ziyue

AU - Guo, Wensheng

PY - 2010/7

Y1 - 2010/7

N2 - In classical smoothing splines, the smoothness is controlled by a single smoothing parameter that penalizes the roughness uniformly across the whole domain. Adaptive smoothing splines extend this framework to allow the smoothing parameter to change in the domain, adapting to the change of roughness. In this article we propose a data driven method to nonparametrically model the penalty function. We propose to approximate the penalty function by a step function whose segmentation is data driven, and to estimate it by maximizing the generalized likelihood. A complexity penalty is added to the generalized likelihood in selecting the best step function from a collection of candidates. A state space representation for the adaptive smoothing splines is derived to ease the computational demand. To allow for fast search among the candidate models, we impose a binary tree structure on the penalty function and propose an efficient search algorithm. We show the consistency of the final estimate. We demonstrate the effectiveness of the method through simulations and a data example.

AB - In classical smoothing splines, the smoothness is controlled by a single smoothing parameter that penalizes the roughness uniformly across the whole domain. Adaptive smoothing splines extend this framework to allow the smoothing parameter to change in the domain, adapting to the change of roughness. In this article we propose a data driven method to nonparametrically model the penalty function. We propose to approximate the penalty function by a step function whose segmentation is data driven, and to estimate it by maximizing the generalized likelihood. A complexity penalty is added to the generalized likelihood in selecting the best step function from a collection of candidates. A state space representation for the adaptive smoothing splines is derived to ease the computational demand. To allow for fast search among the candidate models, we impose a binary tree structure on the penalty function and propose an efficient search algorithm. We show the consistency of the final estimate. We demonstrate the effectiveness of the method through simulations and a data example.

KW - Binary tree

KW - Complexity penalty

KW - Generalized maximum likelihood

KW - Model selection

KW - State space method

UR - http://www.scopus.com/inward/record.url?scp=78349252072&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78349252072&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:78349252072

VL - 20

SP - 1143

EP - 1163

JO - Statistica Sinica

JF - Statistica Sinica

SN - 1017-0405

IS - 3

ER -