Exact Inference Using a Regression/Attraction Method of Markov Chain Monte Carlo



Philip J. O'Neil

AnVil Informatics, Inc.


In recent years much work has been focused on developing exact tests for the analysis of discrete data using log linear or logistic regression models. A parametric model is tested for a data set by conditioning on the value of a sufficient statistic and determining the probability of obtaining another data set as extreme or more extreme relative to the general model, where extremeness is determined by the value of a test statistic. Exact determination of these probabilities can be infeasible for high dimensional problems and asymptotic approximations to them are often inaccurate when there are small data entries and/or there are many nuisance parameters. In these cases Monte Carlo methods can be used to estimate exact probabilities by randomly generating data sets (tables) that match the sufficient statistic of the original table. However, naive Monte Carlo methods produce tables that are usually far from matching the sufficient statistic. The Markov chain Monte Carlo method presented in this talk (the regression/attraction approach) uses attraction to concentrate the distribution around the set of tables that match the sufficient statistic, and uses regression to take advantage of information in tables that ``almost'' match. It is also more general than others in that it does not require the sufficient statistic to be linear, and it can be adapted to problems involving continuous data.

Back to statistics seminar home page