How to perform a Prediction with Matlabs Hidden Markov Model? (Statistics Toolbox)

Question

I have trained a HMM in matlab using the hmmtrain-function from the statistics package.

Using hmmdecode and my given evidence between t_0 and t_1, I can do Filtering and Smoothing to compute the posterior distribution over the possible states between t_0 and t_1.

Now I want to compute the posterior distribution over a future state later than t_1, given all evidence to date. In theroy, this means I have to apply the forward-algorithm on the markov chain.

Do you know which matlab-function I can use to do this? I suppose that there is something built-in in the statistics package, I just cannot find it.

Answer 1

To get the probability over hidden states at t_2, just multiply your posterior over t_1 by your transition matrix.

Answer 2

It turns out the matlab statistics package is a bit scarce when it comes to HMMs. Kevin Murphy's toolbox seems to offer much more functionality and a lot of convenience functions.