Longitudinal functional models with structured penalties

Madan G. Kundu, Jaroslaw Harezlak, Timothy W. Randolph

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

This article addresses estimation in regression models for longitudinally collected functional covariates (time-varying predictor curves) with a longitudinal scalar outcome. The framework consists of estimating a time-varying coefficient function that is modelled as a linear combination of time-invariant functions with time-varying coefficients. The model uses extrinsic information to inform the structure of the penalty, while the estimation procedure exploits the equivalence between penalized least squares estimation and a linear mixed model representation. The process is empirically evaluated with several simulations and it is applied to analyze the neurocognitive impairment of human immunodeficiency virus (HIV) patients and its association with longitudinally-collected magnetic resonance spectroscopy (MRS) curves.

Original languageEnglish (US)
Pages (from-to)114-139
Number of pages26
JournalStatistical Modelling
Volume16
Issue number2
DOIs
StatePublished - 2016

Fingerprint

Time-varying Coefficients
Functional Model
Penalty
Time-varying Covariates
Penalized Least Squares
Linear Mixed Model
Curve
Magnetic Resonance
Least Squares Estimation
Virus
Linear Combination
Spectroscopy
Predictors
Regression Model
Equivalence
Scalar
Invariant
Simulation
Model
Time-varying coefficients

Keywords

  • functional data analysis
  • generalized singular value decomposition
  • Longitudinal data
  • LongPEER estimate
  • structured penalty

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Longitudinal functional models with structured penalties. / Kundu, Madan G.; Harezlak, Jaroslaw; Randolph, Timothy W.

In: Statistical Modelling, Vol. 16, No. 2, 2016, p. 114-139.

Research output: Contribution to journalArticle

Kundu, Madan G. ; Harezlak, Jaroslaw ; Randolph, Timothy W. / Longitudinal functional models with structured penalties. In: Statistical Modelling. 2016 ; Vol. 16, No. 2. pp. 114-139.
@article{7cd59142733d45a79f8760419ca81b2e,
title = "Longitudinal functional models with structured penalties",
abstract = "This article addresses estimation in regression models for longitudinally collected functional covariates (time-varying predictor curves) with a longitudinal scalar outcome. The framework consists of estimating a time-varying coefficient function that is modelled as a linear combination of time-invariant functions with time-varying coefficients. The model uses extrinsic information to inform the structure of the penalty, while the estimation procedure exploits the equivalence between penalized least squares estimation and a linear mixed model representation. The process is empirically evaluated with several simulations and it is applied to analyze the neurocognitive impairment of human immunodeficiency virus (HIV) patients and its association with longitudinally-collected magnetic resonance spectroscopy (MRS) curves.",
keywords = "functional data analysis, generalized singular value decomposition, Longitudinal data, LongPEER estimate, structured penalty",
author = "Kundu, {Madan G.} and Jaroslaw Harezlak and Randolph, {Timothy W.}",
year = "2016",
doi = "10.1177/1471082X15626291",
language = "English (US)",
volume = "16",
pages = "114--139",
journal = "Statistical Modelling",
issn = "1471-082X",
publisher = "SAGE Publications Ltd",
number = "2",

}

TY - JOUR

T1 - Longitudinal functional models with structured penalties

AU - Kundu, Madan G.

AU - Harezlak, Jaroslaw

AU - Randolph, Timothy W.

PY - 2016

Y1 - 2016

N2 - This article addresses estimation in regression models for longitudinally collected functional covariates (time-varying predictor curves) with a longitudinal scalar outcome. The framework consists of estimating a time-varying coefficient function that is modelled as a linear combination of time-invariant functions with time-varying coefficients. The model uses extrinsic information to inform the structure of the penalty, while the estimation procedure exploits the equivalence between penalized least squares estimation and a linear mixed model representation. The process is empirically evaluated with several simulations and it is applied to analyze the neurocognitive impairment of human immunodeficiency virus (HIV) patients and its association with longitudinally-collected magnetic resonance spectroscopy (MRS) curves.

AB - This article addresses estimation in regression models for longitudinally collected functional covariates (time-varying predictor curves) with a longitudinal scalar outcome. The framework consists of estimating a time-varying coefficient function that is modelled as a linear combination of time-invariant functions with time-varying coefficients. The model uses extrinsic information to inform the structure of the penalty, while the estimation procedure exploits the equivalence between penalized least squares estimation and a linear mixed model representation. The process is empirically evaluated with several simulations and it is applied to analyze the neurocognitive impairment of human immunodeficiency virus (HIV) patients and its association with longitudinally-collected magnetic resonance spectroscopy (MRS) curves.

KW - functional data analysis

KW - generalized singular value decomposition

KW - Longitudinal data

KW - LongPEER estimate

KW - structured penalty

UR - http://www.scopus.com/inward/record.url?scp=84962038061&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84962038061&partnerID=8YFLogxK

U2 - 10.1177/1471082X15626291

DO - 10.1177/1471082X15626291

M3 - Article

VL - 16

SP - 114

EP - 139

JO - Statistical Modelling

JF - Statistical Modelling

SN - 1471-082X

IS - 2

ER -