Is the acceptance ratio a measure of the quality of Bayes estimation?

Hi everyone!

Is the acceptance ratio a measure of the quality of Bayes estimation?

If not, what is? MDD?

Thanks!

There are various statistics for various “quality metrics”. What do you have in mind? Model fit? Convergence of the chain?

I’m not sure…I think MDD measures model fit and acceptance ratio measures convergence…Am I right?

MDD measures fit relative to other models on the same data. It is pretty meaningless on its own. The acceptance rate only tells you about efficiency of the MCMC, it does not tell you about convergence. Use e.g. the Geweke diagnostics for that.