Error when using baum-welch for model fitting

[malonzm1]

When I use model.fit(sequence, algorithm='baum_welch') it throws the following error: Traceback (most recent call last): File "", line 1, in File "pomegranate/hmm.pyx", line 2659, in pomegranate.hmm.HiddenMarkovModel.fit File "pomegranate/hmm.pyx", line 2686, in pomegranate.hmm.HiddenMarkovModel.fit TypeError: unsupported operand type(s) for +: 'int' and 'NoneType'

Pls. advise.

[jmschrei]

Please provide more details when opening issues. It's impossible for me to know what's going on here (e.g., did you make an error when defining your model, in the structure of your data, etc), so I can't really provide any useful guidance.

[malonzm1]

Thanks. This is the code I used.

s0 = State(NormalDistribution(0, 0.08), name = 's0') s1 = State(NormalDistribution(0.3, 0.06), name = 's1') s2 = State(NormalDistribution(-0.3, 0.06), name = 's2') model = HiddenMarkovModel() model.add_states(s0, s1, s2) model.add_transition(model.start, s0, 0.333) model.add_transition(model.start, s1, 0.333) model.add_transition(model.start, s2, 0.333) model.add_transition(s0, s0, 0.5) model.add_transition(s0, s1, 0.25) model.add_transition(s0, s2, 0.25) model.add_transition(s1, s0, 0.25) model.add_transition(s1, s1, 0.5) model.add_transition(s1, s2, 0.25) model.add_transition(s2, s0, 0.25) model.add_transition(s2, s1, 0.25) model.add_transition(s2, s2, 0.5) model.add_transition(s0, model.end, 0.333) model.add_transition(s1, model.end, 0.333) model.add_transition(s2, model.end, 0.333) model.bake() model.fit(diffs, algorithm='baum_welch')

It works fine when I use model.fit(sequence,algorithm='viterbi') or when I don't specify the algorithm.

[jmschrei]

The default is baum-welch, which is what you need to pass in, not baum_welch.

[malonzm1]

Thanks!