mlpy.stats.dbn.hmm.GaussianHMM.decode

GaussianHMM.decode(obs, algorithm='viterbi')

Find the most likely state sequence.

Find the most likely state sequence corresponding to the observation obs. Uses the given algorithm for decoding.

Parameters:

obs : array_like, shape (nfeatures, T)

The local evidence vector.

algorithm : {‘viterbi’, ‘map’}

Decoder algorithm to be used.

Returns:

best_path : array_like, shape (n,)

The most likely states for each observation

loglik : float

Log probability of the maximum likelihood path through the HMM