Last updated: Sept. 17, 2020, 9:52 a.m.
Speaker: Prof. Stephen Jun Villejo
Date: Oct. 1, 2020, 5 p.m.
Venue: via ZOOM
The popularity of Bayesian models in spatial epidemiology is increasing because of their flexibility in providing global and local smoothing of estimates. This study proposes the use of a three-level hierarchical model for the incidence of certain diseases in the Philippines using Bayesian inference. Different prior structures, classified according to the extent of the smoothing either local are global, are explored and investigated. Global smoothing may not be valid especially for large and spatially heterogeneous areas. In such cases, local variation in the smoothing, which allows for differential smoothing depending on the neighborhood structure, should be considered. The criteria that will be used for model evaluation are the plausibility of the estimates and goodness-of-fit measures which include the Watanabe-Akaike Information Criterion (WAIC), Deviance Information Criterion (DIC), Moran’s I on the residuals, and visual checks for residual patterns. Computational time and difficulty in the implementation of the models are also assessed. Convergence of the MCMC should also be carefully checked and evaluated. The researcher expects that local smoothing models are more appropriate in modelling diseases in the Philippines due to heterogeneity in the system.
Undersecretary Dr. Lisa Grace S. Bersales, National Statistician and Civil Registrar General of the Philippine Statistics Authority (PSA), was elec...View our Most Read Article
One of the highlights of the 2015 UP School of Statistics Recognition Rites was the announcement of the recipients of the Best Undergraduate Student Paper Award.View our most read journal