[no subject]

[roger gill <mysnews@yahoo.co.uk>]

Dear Listers,

A simple question.  Suppose one, working within a bayesian framework, wished to calculate posterior model probabilities - then we calculate marginals and weight them by the prior for that model.  The resulting calculation is of the following form

a/(a+b)   or b/(a+b)

In this case the marginals are analytically tractable, so a Reversible jump/Carlin and Chib etc need not be used.  The problem is both a and b are too small for Splus 6.1 pro to handle, so how could I go about doing the above two calculations?

Thanks

 

---

Get a bigger mailbox -- choose a size that fits your needs.