Bayesian analysis of age-period-cohort models

This research project addresses two main areas of research. First, methodology on integrated model selection in multivariate Bayesian age-period-cohort models for count data will be discussed and developed. Modern inference techniques, in particular effic ient Markov chain Monte Carlo (MCMC) methods and numerical methods based on integrated nested Laplace approximation will thereby be considered. Secondly, model choice within the integrated nested Laplace approximation (INLA) is addressed. Methodology for a robust estimation of a novel model choice criterion will be investigated and if applicable implemented.