main.py

import hmm as gmove
import numpy as np
from hmmlearn import hmm

# X_example = np.array([[40, -70, 42320, 0],
# 					  [42, -71, 43989, 1],
# 					  [38, -72, 32324, 2],
# 					  [42, -70, 23222, 3],
# 					  [37, -71, 81999, 3]])
length_example = [3, 2]
weight_example = [1/2, 1/2]


X_example = np.array([[0],
					  [0],
					  [0],
					  [1]])

# model = gmove.GroupLevelHMM(n_components=2, init_params='mce')
# model.set_weights(weight_example)
# model.fit(X_example, length_example)

model = hmm.MultinomialHMM(n_components=3)
model.fit(X_example, length_example)
print(model.score(X_example))

# model._init(X_example, length_example)

# obs = X_example[0:3]
# framelogprob = model._compute_log_likelihood(obs)
# logprob, fwdlattice = model._do_forward_pass(framelogprob)
# bwdlattice = model._do_backward_pass(framelogprob)
# posteriors = model._compute_posteriors(fwdlattice, bwdlattice)
# stats = model._initialize_sufficient_statistics()
# model._accumulate_sufficient_statistics(stats, obs, framelogprob, posteriors, fwdlattice, bwdlattice)