IOHMM with missing observations

[aminzadenoori]

I am working on a project that I need to implement the em algorithm for IOHMMs when the transition and observation probabilities are dependant on the inputs that it means P(s2)=p(S1|input,s1) and P(O t)=P(s t-1, input t). Because your code is very well written and well implemented with a little bit modification on your code I can use it in my project.
thank you.