Re: [R] 'all subsets' fitting algorithm for Bayesian approach

2010-10-04 Thread Michael Hopkins
Bert Thanks for your reply but I already have the Box & Steinberg work in my extensive library of 'homework' :o) I have some problem specific priors that I need to use for calculating model probabilities so that I can produce predictive distributions using Bayesian model averaging - hence I

Re: [R] 'all subsets' fitting algorithm for Bayesian approach

2010-10-01 Thread Bert Gunter
u... You are reinventing the wheel. In fact, several wheels: the statistical literature already has several different approaches worked out for this. For example, George Box and David Steinberg did one about 20 years ago, and it has been incorporated as one of the options in the JMP DOE model c