Efficient estimation and prediction for the Bayesian binary spatial model with flexible link functions

Vivekananda Roy, Evangelos Evangelou, Zhengyuan Zhu

Research output: Contribution to journalArticle

7 Citations (Scopus)
94 Downloads (Pure)

Abstract

Spatial generalized linear mixed models (SGLMMs) are popular models for
spatial data with a non-Gaussian response. Binomial SGLMMs with logit or
probit link functions are often used to model spatially dependent
binomial random variables. It is known that for independent Binomial
data, the robit regression model provides a more robust (against extreme
observations) alternative to the more popular logistic and probit models.
In this article, we introduce a Bayesian spatial robit model for
spatially dependent binomial data. Since constructing a meaningful prior
on the link function parameter as well as the spatial correlation
parameters in SGLMMs is difficult, we propose an empirical Bayes (EB)
approach for the estimation of these parameters as well as for the
prediction of the random effects. The EB methodology is implemented by
efficient importance sampling methods based on Markov chain Monte Carlo
(MCMC) algorithms. Our simulation study shows that the robit model is
robust against model misspecification, and our EB method results in
estimates with less bias than full Bayesian (FB) analysis.
The methodology is applied to a Celastrus Orbiculatus data, and a
Rhizoctonia root data. For the former, the robit model is shown to do better
for predicting the spatial distribution of an invasive species which is
known to contain outlying observations. For the latter, our approach is
doing as well as the classical models for predicting the disease severity
for a root disease, as the probit link is shown to be appropriate.

Though this paper is written for Binomial SGLMMs for brevity, the EB
methodology is more general and can be applied to other types of SGLMMs.
In the accompanying R package geoBayes, implementations for other SGLMMs
such as Poisson and Gamma SGLMMs are provided.
Original languageEnglish
Pages (from-to)289-298
JournalBiometrics
Volume72
Issue number1
Early online date31 Aug 2015
DOIs
Publication statusPublished - Mar 2016

Fingerprint

Generalized Linear Mixed Model
Link Function
Bayes Theorem
Efficient Estimation
Spatial Model
Linear Models
Binary
prediction
Prediction
Empirical Bayes
Probit
Methodology
Celastrus
Roots
Empirical Bayes Method
Rhizoctonia
Introduced Species
Probit Model
Markov Chains
Model Misspecification

Cite this

Efficient estimation and prediction for the Bayesian binary spatial model with flexible link functions. / Roy, Vivekananda; Evangelou, Evangelos; Zhu, Zhengyuan.

In: Biometrics, Vol. 72, No. 1, 03.2016, p. 289-298.

Research output: Contribution to journalArticle

@article{ad44ac1555b848c6bd4d17a1e260c380,
title = "Efficient estimation and prediction for the Bayesian binary spatial model with flexible link functions",
abstract = "Spatial generalized linear mixed models (SGLMMs) are popular models for spatial data with a non-Gaussian response. Binomial SGLMMs with logit or probit link functions are often used to model spatially dependent binomial random variables. It is known that for independent Binomial data, the robit regression model provides a more robust (against extreme observations) alternative to the more popular logistic and probit models. In this article, we introduce a Bayesian spatial robit model for spatially dependent binomial data. Since constructing a meaningful prior on the link function parameter as well as the spatial correlation parameters in SGLMMs is difficult, we propose an empirical Bayes (EB) approach for the estimation of these parameters as well as for the prediction of the random effects. The EB methodology is implemented by efficient importance sampling methods based on Markov chain Monte Carlo (MCMC) algorithms. Our simulation study shows that the robit model is robust against model misspecification, and our EB method results in estimates with less bias than full Bayesian (FB) analysis. The methodology is applied to a Celastrus Orbiculatus data, and a Rhizoctonia root data. For the former, the robit model is shown to do better for predicting the spatial distribution of an invasive species which is known to contain outlying observations. For the latter, our approach is doing as well as the classical models for predicting the disease severity for a root disease, as the probit link is shown to be appropriate. Though this paper is written for Binomial SGLMMs for brevity, the EB methodology is more general and can be applied to other types of SGLMMs. In the accompanying R package geoBayes, implementations for other SGLMMs such as Poisson and Gamma SGLMMs are provided.",
author = "Vivekananda Roy and Evangelos Evangelou and Zhengyuan Zhu",
year = "2016",
month = "3",
doi = "10.1111/biom.12371",
language = "English",
volume = "72",
pages = "289--298",
journal = "Biometrics",
issn = "0006-341X",
publisher = "Wiley-Blackwell",
number = "1",

}

TY - JOUR

T1 - Efficient estimation and prediction for the Bayesian binary spatial model with flexible link functions

AU - Roy, Vivekananda

AU - Evangelou, Evangelos

AU - Zhu, Zhengyuan

PY - 2016/3

Y1 - 2016/3

N2 - Spatial generalized linear mixed models (SGLMMs) are popular models for spatial data with a non-Gaussian response. Binomial SGLMMs with logit or probit link functions are often used to model spatially dependent binomial random variables. It is known that for independent Binomial data, the robit regression model provides a more robust (against extreme observations) alternative to the more popular logistic and probit models. In this article, we introduce a Bayesian spatial robit model for spatially dependent binomial data. Since constructing a meaningful prior on the link function parameter as well as the spatial correlation parameters in SGLMMs is difficult, we propose an empirical Bayes (EB) approach for the estimation of these parameters as well as for the prediction of the random effects. The EB methodology is implemented by efficient importance sampling methods based on Markov chain Monte Carlo (MCMC) algorithms. Our simulation study shows that the robit model is robust against model misspecification, and our EB method results in estimates with less bias than full Bayesian (FB) analysis. The methodology is applied to a Celastrus Orbiculatus data, and a Rhizoctonia root data. For the former, the robit model is shown to do better for predicting the spatial distribution of an invasive species which is known to contain outlying observations. For the latter, our approach is doing as well as the classical models for predicting the disease severity for a root disease, as the probit link is shown to be appropriate. Though this paper is written for Binomial SGLMMs for brevity, the EB methodology is more general and can be applied to other types of SGLMMs. In the accompanying R package geoBayes, implementations for other SGLMMs such as Poisson and Gamma SGLMMs are provided.

AB - Spatial generalized linear mixed models (SGLMMs) are popular models for spatial data with a non-Gaussian response. Binomial SGLMMs with logit or probit link functions are often used to model spatially dependent binomial random variables. It is known that for independent Binomial data, the robit regression model provides a more robust (against extreme observations) alternative to the more popular logistic and probit models. In this article, we introduce a Bayesian spatial robit model for spatially dependent binomial data. Since constructing a meaningful prior on the link function parameter as well as the spatial correlation parameters in SGLMMs is difficult, we propose an empirical Bayes (EB) approach for the estimation of these parameters as well as for the prediction of the random effects. The EB methodology is implemented by efficient importance sampling methods based on Markov chain Monte Carlo (MCMC) algorithms. Our simulation study shows that the robit model is robust against model misspecification, and our EB method results in estimates with less bias than full Bayesian (FB) analysis. The methodology is applied to a Celastrus Orbiculatus data, and a Rhizoctonia root data. For the former, the robit model is shown to do better for predicting the spatial distribution of an invasive species which is known to contain outlying observations. For the latter, our approach is doing as well as the classical models for predicting the disease severity for a root disease, as the probit link is shown to be appropriate. Though this paper is written for Binomial SGLMMs for brevity, the EB methodology is more general and can be applied to other types of SGLMMs. In the accompanying R package geoBayes, implementations for other SGLMMs such as Poisson and Gamma SGLMMs are provided.

UR - http://dx.doi.org/10.1111/biom.12371

U2 - 10.1111/biom.12371

DO - 10.1111/biom.12371

M3 - Article

VL - 72

SP - 289

EP - 298

JO - Biometrics

JF - Biometrics

SN - 0006-341X

IS - 1

ER -