Open Conference Systems, 50th Scientific meeting of the Italian Statistical Society

Font Size: 
Assessment of the INLA approach on gerarchic bayesian models for the spatial disease distribution: a real data application
Paolo Girardi, Emanuela Bovo, Carmen Stocco, Susanna Baracco, Alberto Rosano, Daniele Monetti, Silvia Rizzato, Sara Zamberlan, Enrico Chinellato, Ugo Fedeli, Massimo Rugge

Last modified: 2018-05-10

Abstract


The use of Laplace approximation methods as INLA (Integrated Nested Laplace Approximation) is being widely used in Bayesian inference, especially in spatial risk regression as the Besag-York-Molliè (BYM) model. INLA is time saving compared to Monte Carlo simulations based on Markov Chains (MCMC), but it produces differences in estimates \cite{Smedt2015,Carroll2015}. In order to perform a comparison, we consider data from the Veneto Cancer Registry. INLA returns estimates comparable to MCMC, but it appears sensitive to the a-priori distribution, while MCMC estimation seems to be more robust.  INLA is a fast and efficient method. However, particular care must be paid to the choice of the parameter relating to the a-priori distribution in both methods.

References


1. De Smedt, T., Simons, K., Van Nieuwenhuyse, A., Molenberghs, G. (2015). ComparingMCMCand INLA for disease mapping with Bayesian hierarchical models. Archives of PublicHealth, 73(1), O2.

2. Carroll, R., Lawson, A. B., Faes, C., Kirby, R. S., Aregay, M., Watjou, K. (2015). ComparingINLA and OpenBUGS for hierarchical Poisson modeling in disease mapping. Spatial andspatio-temporal epidemiology, 14, 45-54.

3. Rue, H., Martino, S., Chopin, N. (2009). Approximate Bayesian inference for latent Gaussianmodels by using integrated nested Laplace approximations. Journal of the royal statisticalsociety: Series b (statistical methodology), 71(2), 319-392.

4. Besag, J., York, J., Mollie, A. Bayesian image restoration with two applications in spatialstatistics (with discussion) Ann Inst Stat Math. 1991; 43: 1–59. doi: 10.1007. BF00116466

5. Lindgren, F., Rue, H. (2015). Bayesian spatial modelling with R-INLA. Journal of StatisticalSoftware, 63(19).

6. Blangiardo, M., Cameletti, M., Baio, G., Rue, H. (2013). Spatial and spatio-temporal modelswith R-INLA. Spatial and spatio-temporal epidemiology, 4, 33-49.

7. Bilancia, M., Demarinis, G. (2014). Bayesian scanning of spatial disease rates with integratednested Laplace approximation (INLA). Statistical Methods & Applications, 23(1), 71-94.


Full Text: PDF