Reisensburg 1998: Abstract Plummer
Statistical Computing '98 - Schloß Reisensburg
Introduction to BUGS
International Agency for Research on Cancer
Markov Chain Monte Carlo (MCMC) simulation techniques have
revolutionized practical Bayesian statistics by making a wide
range of complex statistical models accessible to analysis.
BUGS (Bayesian inference Using Gibbs Sampling) is a program
which provides a language for expressing complex statistical
models and an MCMC sampler which produces the simulations.
This tutorial will introduce BUGS and illustrate some of the
problems it is suitable for.
Topics covered will include:
- Introduction to graphical models and Markov Chain Monte
- Description of the BUGS language and commands.
Convergence diagnostics and output analysis using the CODA package.
- Application to problems of missing data, measurement error,
random effects, ...
Exercises will be provided for you to try on your own PC, so come
prepared. BUGS and CODA can be downloaded from the MRC Biostatistics
Unit Web site:
The CODA package runs on top of S-PLUS. If you do not have an S-PLUS license you can use the
freely available R language
and the port of CODA to R
- Gilks WR, Clayton DG, Spiegelhalter DJ, Best NG, McNeil AJ,
Sharples LD and Kirby AJ. Modelling complexity: applications
of Gibbs sampling in medicine. J R Stat Soc B (1993), 1993, 55,
No 1, pp 39-52.
- Smith, AFM and Roberts GO. Bayesian computation via the Gibbs
sampler and related Markov chain Monte Carlo methods.
J R Stat Soc, B (1993), 55, No 1, pp 3-23.
30. Statistical Computing '98 ---