Bayesian analysis of spatial generalized linear mixed models with Laplace moving average random fields

Adam Walder, Ephraim M. Hanks

Research output: Contribution to journalArticle

Abstract

Gaussian random field (GRF) models are widely used in spatial statistics to capture spatially correlated error. Gaussian processes can easily be replaced by the less commonly used Laplace moving averages (LMAs) in spatial generalized linear mixed models (SGLMMs). LMAs are shown to offer improved predictive power when the data exhibits localized spikes in the response. Further, SGLMMs with LMAs are shown to maintain analogous parameter inference and similar computing to Gaussian SGLMMs. A novel discrete space LMA model for irregular lattices is proposed, along with conjugate samplers for LMAs with georeferenced and areal support. A Bayesian analysis of SGLMMs with LMAs and GRFs is conducted over multiple data support and response types.

Original languageEnglish (US)
Article number106861
JournalComputational Statistics and Data Analysis
Volume144
DOIs
StatePublished - Apr 2020

Fingerprint

Generalized Linear Mixed Model
Moving Average
Bayesian Analysis
Laplace
Random Field
Spatial Statistics
Correlated Errors
Moving Average Model
Gaussian Random Field
Spike
Gaussian Process
Irregular
Statistics
Computing

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Computational Mathematics
  • Computational Theory and Mathematics
  • Applied Mathematics

Cite this

@article{a62cd6a3dc4f4ed9be7ba1aefcc947b9,
title = "Bayesian analysis of spatial generalized linear mixed models with Laplace moving average random fields",
abstract = "Gaussian random field (GRF) models are widely used in spatial statistics to capture spatially correlated error. Gaussian processes can easily be replaced by the less commonly used Laplace moving averages (LMAs) in spatial generalized linear mixed models (SGLMMs). LMAs are shown to offer improved predictive power when the data exhibits localized spikes in the response. Further, SGLMMs with LMAs are shown to maintain analogous parameter inference and similar computing to Gaussian SGLMMs. A novel discrete space LMA model for irregular lattices is proposed, along with conjugate samplers for LMAs with georeferenced and areal support. A Bayesian analysis of SGLMMs with LMAs and GRFs is conducted over multiple data support and response types.",
author = "Adam Walder and Hanks, {Ephraim M.}",
year = "2020",
month = "4",
doi = "10.1016/j.csda.2019.106861",
language = "English (US)",
volume = "144",
journal = "Computational Statistics and Data Analysis",
issn = "0167-9473",
publisher = "Elsevier",

}

TY - JOUR

T1 - Bayesian analysis of spatial generalized linear mixed models with Laplace moving average random fields

AU - Walder, Adam

AU - Hanks, Ephraim M.

PY - 2020/4

Y1 - 2020/4

N2 - Gaussian random field (GRF) models are widely used in spatial statistics to capture spatially correlated error. Gaussian processes can easily be replaced by the less commonly used Laplace moving averages (LMAs) in spatial generalized linear mixed models (SGLMMs). LMAs are shown to offer improved predictive power when the data exhibits localized spikes in the response. Further, SGLMMs with LMAs are shown to maintain analogous parameter inference and similar computing to Gaussian SGLMMs. A novel discrete space LMA model for irregular lattices is proposed, along with conjugate samplers for LMAs with georeferenced and areal support. A Bayesian analysis of SGLMMs with LMAs and GRFs is conducted over multiple data support and response types.

AB - Gaussian random field (GRF) models are widely used in spatial statistics to capture spatially correlated error. Gaussian processes can easily be replaced by the less commonly used Laplace moving averages (LMAs) in spatial generalized linear mixed models (SGLMMs). LMAs are shown to offer improved predictive power when the data exhibits localized spikes in the response. Further, SGLMMs with LMAs are shown to maintain analogous parameter inference and similar computing to Gaussian SGLMMs. A novel discrete space LMA model for irregular lattices is proposed, along with conjugate samplers for LMAs with georeferenced and areal support. A Bayesian analysis of SGLMMs with LMAs and GRFs is conducted over multiple data support and response types.

UR - http://www.scopus.com/inward/record.url?scp=85074151815&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85074151815&partnerID=8YFLogxK

U2 - 10.1016/j.csda.2019.106861

DO - 10.1016/j.csda.2019.106861

M3 - Article

AN - SCOPUS:85074151815

VL - 144

JO - Computational Statistics and Data Analysis

JF - Computational Statistics and Data Analysis

SN - 0167-9473

M1 - 106861

ER -