Just suggesting two simple enhancements. 1) adding variable specifying which latent class the probability.scale parameter belongs to, and 2) number of latent classes in summaries. Both should be easy enhancements (though admittedly not important as user can add manually). For 1), I suggest a simple rep(1:n_classes, each=n_unique_parameters*n_categories), which is logically faulty but assumption is pretty safe.
Keep up great work!
Just suggesting two simple enhancements. 1) adding variable specifying which latent class the probability.scale parameter belongs to, and 2) number of latent classes in summaries. Both should be easy enhancements (though admittedly not important as user can add manually). For 1), I suggest a simple rep(1:n_classes, each=n_unique_parameters*n_categories), which is logically faulty but assumption is pretty safe.
Keep up great work!