Bayes Rules Cannot Be Beat!


Tutorial Notes and Code

Visitors are encouraged to download my course notes on Bayesian inference and Markov chain Monte Carlo.  I taught my first PhD course on Bayesian inference in 1985, and have updated and revised my notes over the years.  They also form the basis of workshops that I have presented to professional statisticians and marketing researchers.  Many researchers have found them to be a good reference for mathematical details.

The notes are split between pdfs.  The first file, in addition to summarizing the highlights of Bayesian theory, develops the distribution theory and matrix algebra to analyze some standard models, such as hierarchical Bayes multivariate regression, logit and probit models.  The notes are in "pseudo" book style and walks through the algebra of Bayesian inference.  The pdf file is very large and takes a few moments to download.  Also, the first two pages are blank - a feature of Tex's book document class.

Bayes Notes

I have implemented these procedures in GAUSS.  The second pdf contains the GAUSS code for these procedures, both to simulate data from the model and to analyze the data via MCMC.  Students of Bayesian inference and WinBugs users may be interested in the code to better understand the algorithms.  A third resource is a zip file of the code. 

Program Notes



Bayes Intro is an introductory lecture on Bayes.

SCECR 2007 Tutorial

Variable Importance are some thoughts about that topic,  presented at GfK's IMSM 2005 conference

Absent Dimensions estimates multivariate normal models when entire dimensions are absent or unobserved