Sacha Morin

Thanks for reporting this. 1. Can you check the observations from classes 1,2,4,5,6,9? Specifically, are they identical or extremely similar? 1. Have you tried fitting an estimator with fewer classes?...

Given that the 6 classes are identical in terms of parameters, you should see very similar probabilities in `predict_proba` for the observations that get assigned to class 6. I suspect...

By other settings name, do you mean you’re shuffling the order of the variables? Shuffling should not affect the fit quality of the overall model, but could affect the order...

Looking at your previous results, the clusterings still look good. Each cluster captures a class (or a part of it if you have more clusters than classes). It's also possible...

Closing. Feel free to reopen if you still want to discuss.

We should also detail what to do with convergence warnings.

@ericlacourse @FelixLaliberte I think we discussed this already and it has been requested by users a few times. I don't have cycles to take care of this these days, but...

@yuanjames are you still stuck with this? I will close, but feel free to reopen if needed.

Thanks for the suggestions! Sample-size adjusted BIC and CAIC seem easy enough. I'm not 100% about BLRT but I would like to hear StepMix power user @FelixLaliberte's thoughts on this....

Reviewing [this paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3904803/#R26), it seems the suggested Sample-Adjusted BIC was missing an `X.shape[0]` in the log. The current suggested implementation is ```python3 n = X.shape[0] return -2 * self.score(X, Y)...