Extending distributed lag models to higher degrees

Matthew J. Heaton, Roger Peng

Research output: Contribution to journalArticle

Abstract

Distributed lag (DL) models relate lagged covariates to a response and are a popular statistical model used in a wide variety of disciplines to analyze exposure-response data. However, classical DL models do not account for possible interactions between lagged predictors. In the presence of interactions between lagged covariates, the total effect of a change on the response is not merely a sum of lagged effects as is typically assumed. This article proposes a new class of models, called high-degree DL models, that extend basic DL models to incorporate hypothesized interactions between lagged predictors. The modeling strategy utilizes Gaussian processes to counterbalance predictor collinearity and as a dimension reduction tool. To choose the degree and maximum lags used within the models, a computationally manageable model comparison method is proposed based on maximum a posteriori estimators. The models and methods are illustrated via simulation and application to investigating the effect of heat exposure on mortality in Los Angeles and New York.

Original languageEnglish (US)
Pages (from-to)398-412
Number of pages15
JournalBiostatistics
Volume15
Issue number2
DOIs
StatePublished - Apr 2014

Fingerprint

Los Angeles
Statistical Models
Hot Temperature
Predictors
Mortality
Covariates
Model
Interaction
Collinearity
Comparison Method
Model Comparison
Maximum a Posteriori
Dimension Reduction
Distributed lag model
Gaussian Process
Statistical Model
Heat
Choose
Estimator
Modeling

Keywords

  • Dimension reduction
  • Gaussian process
  • Heat exposure
  • Lagged interaction
  • NMMAPS dataset

ASJC Scopus subject areas

  • Medicine(all)
  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Extending distributed lag models to higher degrees. / Heaton, Matthew J.; Peng, Roger.

In: Biostatistics, Vol. 15, No. 2, 04.2014, p. 398-412.

Research output: Contribution to journalArticle

Heaton, Matthew J. ; Peng, Roger. / Extending distributed lag models to higher degrees. In: Biostatistics. 2014 ; Vol. 15, No. 2. pp. 398-412.
@article{0f94c825d7cb4cba92c235a9ec4bf904,
title = "Extending distributed lag models to higher degrees",
abstract = "Distributed lag (DL) models relate lagged covariates to a response and are a popular statistical model used in a wide variety of disciplines to analyze exposure-response data. However, classical DL models do not account for possible interactions between lagged predictors. In the presence of interactions between lagged covariates, the total effect of a change on the response is not merely a sum of lagged effects as is typically assumed. This article proposes a new class of models, called high-degree DL models, that extend basic DL models to incorporate hypothesized interactions between lagged predictors. The modeling strategy utilizes Gaussian processes to counterbalance predictor collinearity and as a dimension reduction tool. To choose the degree and maximum lags used within the models, a computationally manageable model comparison method is proposed based on maximum a posteriori estimators. The models and methods are illustrated via simulation and application to investigating the effect of heat exposure on mortality in Los Angeles and New York.",
keywords = "Dimension reduction, Gaussian process, Heat exposure, Lagged interaction, NMMAPS dataset",
author = "Heaton, {Matthew J.} and Roger Peng",
year = "2014",
month = "4",
doi = "10.1093/biostatistics/kxt031",
language = "English (US)",
volume = "15",
pages = "398--412",
journal = "Biostatistics",
issn = "1465-4644",
publisher = "Oxford University Press",
number = "2",

}

TY - JOUR

T1 - Extending distributed lag models to higher degrees

AU - Heaton, Matthew J.

AU - Peng, Roger

PY - 2014/4

Y1 - 2014/4

N2 - Distributed lag (DL) models relate lagged covariates to a response and are a popular statistical model used in a wide variety of disciplines to analyze exposure-response data. However, classical DL models do not account for possible interactions between lagged predictors. In the presence of interactions between lagged covariates, the total effect of a change on the response is not merely a sum of lagged effects as is typically assumed. This article proposes a new class of models, called high-degree DL models, that extend basic DL models to incorporate hypothesized interactions between lagged predictors. The modeling strategy utilizes Gaussian processes to counterbalance predictor collinearity and as a dimension reduction tool. To choose the degree and maximum lags used within the models, a computationally manageable model comparison method is proposed based on maximum a posteriori estimators. The models and methods are illustrated via simulation and application to investigating the effect of heat exposure on mortality in Los Angeles and New York.

AB - Distributed lag (DL) models relate lagged covariates to a response and are a popular statistical model used in a wide variety of disciplines to analyze exposure-response data. However, classical DL models do not account for possible interactions between lagged predictors. In the presence of interactions between lagged covariates, the total effect of a change on the response is not merely a sum of lagged effects as is typically assumed. This article proposes a new class of models, called high-degree DL models, that extend basic DL models to incorporate hypothesized interactions between lagged predictors. The modeling strategy utilizes Gaussian processes to counterbalance predictor collinearity and as a dimension reduction tool. To choose the degree and maximum lags used within the models, a computationally manageable model comparison method is proposed based on maximum a posteriori estimators. The models and methods are illustrated via simulation and application to investigating the effect of heat exposure on mortality in Los Angeles and New York.

KW - Dimension reduction

KW - Gaussian process

KW - Heat exposure

KW - Lagged interaction

KW - NMMAPS dataset

UR - http://www.scopus.com/inward/record.url?scp=84896786506&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896786506&partnerID=8YFLogxK

U2 - 10.1093/biostatistics/kxt031

DO - 10.1093/biostatistics/kxt031

M3 - Article

C2 - 23990524

AN - SCOPUS:84896786506

VL - 15

SP - 398

EP - 412

JO - Biostatistics

JF - Biostatistics

SN - 1465-4644

IS - 2

ER -