🚀 go-pugleaf

         DEPARTMENT OF COMPUTER SCIENCE
           COLORADO STATE UNIVERSITY

                  BARNEY'S
        MONDAY AFTERNOON CLUB SEMINAR

	Speaker:	Jennifer Hoeting
	                Department of Statistics
	                Colorado State University
	   
	   Date:	October 7, 1996
	   Time:	3:30 - 4:30 p.m.
	  Place:	100 USC

       Refreshments:  following the talk

      BAYESIAN MODEL AVERAGING FOR LINEAR 
               REGRESSION MODELS

I consider the problem of accounting for model 
uncertainty in linear regression models, specifically 
addressing the uncertainty involved in variable 
selection, outlier identification, and transformation 
selection.  The standard practice of conditioning 
on a single selected model ignores model uncertainty 
which leads to the underestimation of uncertainty in 
subsequent inferences.  A Bayesian solution to this 
problem involves averaging over all possible models 
when making inferences about quantities of interest.  
However, this approach is often not practical due to 
the large number of models under consideration.  I will 
describe a Markov chain Monte Carlo (MCMC) approach 
which directly approximates the exact model average.  
In this context MCMC allows for a probabilistic search 
of a complex model space.  Model averaging procedures 
provide better predictive performance than any single 
model which might reasonably have been selected.