Jacob Schreiber

Yes, you should be able to use `IndependentComponents`, which treats each feature independently with a different distribution: https://github.com/jmschrei/pomegranate/blob/master/pomegranate/distributions/independent_components.py#L15 A downside of this is that it does treat the two features...

Howdy. When you have a dense transition matrix you can pass that in to `edges`. See the section "Dense and Sparse HMMs": https://github.com/jmschrei/pomegranate/blob/master/docs/tutorials/B_Model_Tutorial_4_Hidden_Markov_Models.ipynb Is that not sufficient?

`predict_proba` gives you the posterior probabilities that each observation aligns to each hidden state in the model *given* all of the other observations in the sequence. It's also called the...

I think you're entering one of the confusing areas of probability theory. Basically, just because a point estimate is above 1 doesn't mean that it's guaranteed to happen. For instance,...

1. Having positive log probability values isn't a problem that needs fixing. The math is still all valid, one just needs to know what it means and why. 2. If...

Hi @Koenig128. Sorry for the delay on this. It turns out that there are a series of small bugs that sometimes mask each other. I am working my way through...

This should be fixed in v1.0.4 and I've added in a unit test with this as an example. Please let me know if you encounter any other issues.

If you pass in `inertia=1.0` to your HMM you should be able to freeze the transition matrix while keeping the emissions unfrozen.

I think that `ends` should be derived from `T` not `E` right? Not sure where you're getting the first call for `P0` but, assuming it's right, I think what you...

I don't think BW here is dependent on the end states but having it can add constraints that make optimization easier. It's hard to diagnose why one might observe the...