Hidden Markov Model
Are there any plan to implement a Hidden Markov Model module that can handle vector observation timeseries?
You are very welcome. It was a lot of work, but should be reasonable fast *and* memory efficient. Plus you can read all the code, since it's a complete implementation. May I ask what you're using HMM's for?
Did you review the following? http://www.sagemath.org/doc/reference/sage/stats/hmm/hmm.html Can you contrast a bit more? Not everyone has domain expertise in time-series.