Looking into Spatio-Temporal Models in Disease Mapping

• Activity Number: 266
• Type: Invited
• Date/Time: Tuesday, August 5, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Lola Ugarte*+ and Miguel Angel Martinez-Beneito and Aritz Adín and Paloma Botella-Rocamora and Tomás Goicoa and Ana F. Militino
• Companies: Public University of Navarre and Generalitat Valenciana and Public University of Navarre and Universidad CEU-Cardenal Herrera and Public University of Navarre and Public University of Navarre
• Keywords: brain cancer incidence ; CAR models ; spatial moving average models ; space-time interactions ; MCMC ; INLA

Abstract

In this talk we plan to compare different spatio-temporal models for smoothing risks in disease mapping. These models will consider alternative ways of introducing both spatial and temporal dependence as well as spatio-temporal interactions. New proposals like spatio-temporal moving average risk smoothing models (STMARS) or autoregressive approaches will be compared with well-known spatio-temporal models including space-time interactions specified with Kronecker products. Brain cancer incidence data will be used to illustrate the behavior of the different models. All the models will be fitted under a fully Bayes approach.