Unified inference for sparse and dense longitudinal models

Seonjin Kim, Zhibiao Zhao

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

In longitudinal data analysis, statistical inference for sparse data and dense data could be substantially different. For kernel smoothing, the estimate of the mean function, the convergence rates and the limiting variance functions are different in the two scenarios. This phenomenon poses challenges for statistical inference, as a subjective choice between the sparse and dense cases may lead to wrong conclusions. We develop methods based on self-normalization that can adapt to the sparse and dense cases in a unified framework. Simulations show that the proposed methods outperform some existing methods.

Original languageEnglish (US)
Pages (from-to)203-212
Number of pages10
JournalBiometrika
Volume100
Issue number1
DOIs
StatePublished - Mar 1 2013

Fingerprint

Statistical Inference
Self-normalization
Longitudinal Data Analysis
Kernel Smoothing
Variance Function
Sparse Data
Convergence Rate
data analysis
Limiting
methodology
Model
Scenarios
seeds
Estimate
Inference
Simulation
Statistical inference
Statistical Data Interpretation
Framework
Normalization

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Mathematics(all)
  • Agricultural and Biological Sciences (miscellaneous)
  • Agricultural and Biological Sciences(all)
  • Statistics, Probability and Uncertainty
  • Applied Mathematics

Cite this

Kim, Seonjin ; Zhao, Zhibiao. / Unified inference for sparse and dense longitudinal models. In: Biometrika. 2013 ; Vol. 100, No. 1. pp. 203-212.
@article{1a26567bce3f445789c84be6dda0762a,
title = "Unified inference for sparse and dense longitudinal models",
abstract = "In longitudinal data analysis, statistical inference for sparse data and dense data could be substantially different. For kernel smoothing, the estimate of the mean function, the convergence rates and the limiting variance functions are different in the two scenarios. This phenomenon poses challenges for statistical inference, as a subjective choice between the sparse and dense cases may lead to wrong conclusions. We develop methods based on self-normalization that can adapt to the sparse and dense cases in a unified framework. Simulations show that the proposed methods outperform some existing methods.",
author = "Seonjin Kim and Zhibiao Zhao",
year = "2013",
month = "3",
day = "1",
doi = "10.1093/biomet/ass050",
language = "English (US)",
volume = "100",
pages = "203--212",
journal = "Biometrika",
issn = "0006-3444",
publisher = "Oxford University Press",
number = "1",

}

Unified inference for sparse and dense longitudinal models. / Kim, Seonjin; Zhao, Zhibiao.

In: Biometrika, Vol. 100, No. 1, 01.03.2013, p. 203-212.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Unified inference for sparse and dense longitudinal models

AU - Kim, Seonjin

AU - Zhao, Zhibiao

PY - 2013/3/1

Y1 - 2013/3/1

N2 - In longitudinal data analysis, statistical inference for sparse data and dense data could be substantially different. For kernel smoothing, the estimate of the mean function, the convergence rates and the limiting variance functions are different in the two scenarios. This phenomenon poses challenges for statistical inference, as a subjective choice between the sparse and dense cases may lead to wrong conclusions. We develop methods based on self-normalization that can adapt to the sparse and dense cases in a unified framework. Simulations show that the proposed methods outperform some existing methods.

AB - In longitudinal data analysis, statistical inference for sparse data and dense data could be substantially different. For kernel smoothing, the estimate of the mean function, the convergence rates and the limiting variance functions are different in the two scenarios. This phenomenon poses challenges for statistical inference, as a subjective choice between the sparse and dense cases may lead to wrong conclusions. We develop methods based on self-normalization that can adapt to the sparse and dense cases in a unified framework. Simulations show that the proposed methods outperform some existing methods.

UR - http://www.scopus.com/inward/record.url?scp=84874771172&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84874771172&partnerID=8YFLogxK

U2 - 10.1093/biomet/ass050

DO - 10.1093/biomet/ass050

M3 - Article

AN - SCOPUS:84874771172

VL - 100

SP - 203

EP - 212

JO - Biometrika

JF - Biometrika

SN - 0006-3444

IS - 1

ER -