04-30-2016 01:56 PM
Is it possible to compare different numbers of components in a GMM estimated by the Bayesian appraoch in proc fmm? The common way using "kmin/kmax" doesn't work.
Many thanks in advance.
07-28-2016 10:56 AM
Can you give more info on "doesn't work"? Anything in the log, or output, that might indicate what is going on?