Abstract:
|
This day-long short course will introduce participants to modern Bayesian computational methods, with a focus on the wide variety of new and old Markov chain Monte Carlo (MCMC) algorithms. The course will combine discussion of the methods and the principles underlying them with demonstration of the methods. The course will discuss computations for model fitting, model selection, model assessment, and MCMC assessment. There will be a heavy emphasis on hands-on implementation of algorithms, primarily within the NIMBLE platform (r-nimble.org), an R-based system that allows for flexible use of a variety of MCMC approaches, as well as flexible programming of new algorithms that can be applied to hierarchical models specified in the BUGS language.
|