
Calculates the crossentropy between the selected HMM model and the HMMObservationSequence.
The crossentropy is a useful upper bound for the entropy of a model. An approximation to the crossentropy for a model on a observation sequence O of length N is:
H(O) = 1/N log p(O), 
where p(O) is the probability of the observation sequence given the model.
© djmw, October 17, 2010