A1060
Title: A Bayesian spatiotemporal Poisson auto-regressive model for the disease infection rate
Authors: Pierfrancesco Alaimo Di Loro - LUMSA University (Italy) [presenting]
Sujit Sahu - University of Southampton (United Kingdom)
Dankmar Boehning - University of Southampton (United Kingdom)
Abstract: The COVID-19 pandemic provided new modelling challenges to investigate epidemic processes. Poisson auto-regression is extended to incorporate spatiotemporal dependence and characterize the local dynamics by borrowing information from adjacent areas. Adopted in a fully Bayesian framework and implemented through a novel sparse-matrix representation in Stan, the model has been validated through a simulation study. It is used to analyze the weekly COVID-19 cases in the English local authority districts and verify some of the epidemic-driving factors. The model detects substantial spatiotemporal heterogeneity and enables the formalization of novel model-based investigation methods for assessing additional aspects of disease epidemiology.
