Bayesian Spatial Conditional Overdispersion ModelsApplication to infant mortality
- Mabel Morales Otero 1
- Vicente Núñez-Antón 1
- 1 Department of Econometrics and Statistics, University of the Basque Country UPV/EHU, Bilbao, Spain
- Itziar Irigoien (ed. lit.)
- Dae-Jin Lee (ed. lit.)
- Joaquín Martínez-Minaya (ed. lit.)
- María Xosé Rodríguez- Álvarez (ed. lit.)
Publisher: Servicio Editorial = Argitalpen Zerbitzua ; Universidad del País Vasco = Euskal Herriko Unibertsitatea
ISBN: 978-84-1319-267-3
Year of publication: 2020
Pages: 374-377
Congress: International Workshop on Statistical Modelling (35. 2020. Bilbao)
Type: Conference paper
Abstract
In this work we revise Bayesian generalized conditional models for spatial count data with overdispersion. We show their usefulness by tting them to infant mortality rates from Colombian regions. These models assume that the overdispersion present in the data may be caused partially from the spatial dependence that exists among the spatial units. Therefore, regression structures are specied both for the conditional mean and for the dispersion parameter, including also spatial neighborhood structures in the model. We work on the case of spatial count data which follow a Poisson distribution, and focus our attention on the spatial generalized conditional normal Poisson model. Models have been tted with the use of the Markov Chain Monte Carlo (MCMC) algorithms within the context of Bayesian estimation methods.