Reference material for hidden Markov models.

 

In class we'll use Andrew Moore's slides from cmu. MooreSlideshmm14.pdf   and a paper put together David Harte to support the R package HiddenMarkov  HMPackageNotes.pdf .  

 

In addition to these there are several references you might find helpful.  The first is a tutorial by Lawrence Rabiner RabinerTutorial10.1.1.131.2084.pdf (a very well-known figure in information theory and signal processing, coding, etc.).  The next is the original paper describing the "Baum Welch" algorithm for implementing EM for hidden Markov models baum.pdf .  The last is leroux-1992.pdf  who considers maximum likelihood estimation for hmm and develops rigorous conditions for things like identifiability.