I just noticed that it is possible to fit mixture models with brms. But is it possible to define the mixtures in repeated measures setting so that all observations of particular subject belong to same mixture? For example say that I have ids 1-10 and each id has n measurements and I have two mixture components. Can brms estimate the probabilities for each id belonging to each mixture component, or will it necessarily work on individual rows of the data?
That’s currently not possible in brms as mixtures are currently working only on individual rows.
Thanks for quick answer. I’ll try to extract the stan code from brms fit and see if I can modify it accordingly.