Intermediate Bayesian Data Analysis Using WinBUGS and BRugs
*Bradley P. Carlin, University of Minnesota
Keywords: WinBUGS, R, BRugs, joint modeling
We offer a hands-on opportunity to explore the use of WinBUGS, the leading Bayesian software package, in a variety of important models, with special attention paid to joint models for longitudinal measurements and survival data. Model building, MCMC convergence diagnosis and acceleration, and posterior plotting and summarization will be mentioned, as well as important data-analytic procedures such as residual analysis, model adequacy (through Bayesian p-values and CPO statistics), variable selection, and model choice (through posterior probabilities and DIC statistics). In addition to WinBUGS, we will also provide a brief introduction to BRugs, the new version of BUGS available directly within the popular R package, which enables simultaneous use of the features of both languages. Students are invited to bring their own laptops (with WinBUGS and R already installed) to the workshop.