How can I specify a prior in proc fmm using the Bayes statement?

[DaLack]

Many thanks for the answer.

But is there a way to force SAS to accept another prior than the distribution selected in the model statement? Because otherwise I couldn't use the beta distribution for a mixture of normal distributions at least with respect to the option mupriorparms.

[Rick_SAS · Accepted solution]

I've told you as much as I was able to gleen from the doc.  I am not a Bayesian myself, so I am not the right person to advise you about how to define a prior.  I've only used the non-Bayesian portions of PROC FMM.

If I were to guess, I would guess that FMM is limited to the models and priors that are defined in the doc.  The BAYES statement is not a replacement for the power of PROC MCMC for doing nonstandard, general, Bayesian analyses.