Department of Statistical Science
Duke University
presents:
Peter J. Green
P.J.Green@bristol.ac.uk
University of Bristol, UK
"Inference on an Unknown Number of Unknowns: Change Point and Mixture Estimation"
Abstract: In many interesting statistical problems the dimension of the object of inference is not fixed. I will discuss the formulation of such problems using a Bayesian hierarchical model, and develop Markov chain Monte Carlo methods that are capable of traversing the resulting complex parameter space.
Using illustrations from mixture analysis and inference on change-points, I will go on to discuss how to draw useful inferences from the complex posterior distributions that arise. Some comparisons will be drawn with methods based on Dirichlet process priors.
April 4, 1997
4:00 pm - 5:00 pm
116 Old Chem Building Any questions concerning the seminar may be addressed to Cheryl McGhee @ [919] 684-8029 or e-mail cheryl@stat.duke.edu. Please contact the author(s) directly for reprints etc.