Online Program Home
My Program

Abstract Details

Activity Number: 171
Type: Contributed
Date/Time: Monday, August 1, 2016 : 10:30 AM to 12:20 PM
Sponsor: Section on Bayesian Statistical Science
Abstract #320637 View Presentation
Title: Univariate and Joint Spatio-Temporal Models for Predicting the Occurrence of Culicoides Across Belgium
Author(s): Yimer Wasihun Kifle* and Christel Faes and Niel Hens
Companies: Hasselt University and Hasselt University and Hasselt University
Keywords: Hierarchical Bayesian Modeling Framework ; Integrated Nested Laplace Approximation ; Stochastic Partial Differential Equation ; Gaussian Markov Random Fields ; Gaussian Random Fields

Obsoletus complex, dewulfi and chiopterus species of Culicoides have been considered as putative vectors for the transmission of ruminant-animal viruses. The joint spatiotemporal occurrence of these species was not well studied before. This paper proposed univariate and joint spatiotemporal models for predicting their prevalence across Belgium. Estimation, inference and prediction were carried out in the Bayesian hierarchical modeling framework using the integrated nested Laplace approximation and stochastic partial differential equation approaches. We compared the fit and complexity of several models using predictive accuracy measures. Our results confirmed that the prediction accuracy was better in coupled spatiotemporal models (with spacetime interaction) than additive spatiotemporal models (without spacetime interaction). Around summer, the Northeastern and the central part of Belgium had the highest prevalence. However, around winter, these areas in particular and the rest of Belgium in general had the lowest prevalence. Therefore, we conclude that the months of January-March, November and December can be defined as months with the lowest prevalence according to our study.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

Copyright © American Statistical Association