def _exact_loglikelihood(self, ob):
log_transmat = np.zeros((self.n_chains, self.n_states, self.n_states))
log_startprob = np.zeros((self.n_chains, self.n_states))
for idx, chain in enumerate(self.chains_):
log_transmat[idx] = chain._log_transmat
log_startprob[idx] = chain._log_startprob
n_state_combinations = self.n_states ** self.n_chains
state_combinations = [tuple(x) for x in list(itertools.product(np.arange(self.n_states), repeat=self.n_chains))]
n_observations = ob.shape[0]
n_features = ob.shape[1]
fwdlattice = np.zeros((n_observations, n_state_combinations))
# Calculate means and covariances for all state combinations and calculate emission probabilities
weight = (1.0 / float(self.n_chains))
weight_squared = weight * weight
covars = np.zeros((n_state_combinations, n_features)) # TODO: add support for all covariance types
means = np.zeros((n_state_combinations, n_features))
for idx, state_combination in enumerate(state_combinations):
for chain_idx, state in enumerate(state_combination):
chain = self.chains_[chain_idx]
covars[idx] += chain._covars_[state]
means[idx] += chain._means_[state]
covars[idx] *= weight_squared
means[idx] *= weight
framelogprob = log_multivariate_normal_density(ob, means, covars, covariance_type='diag') # TODO: add support for all covariance types
# Run the forward algorithm
fhmmc._forward(n_observations, self.n_chains, self.n_states, state_combinations, log_startprob, log_transmat,
framelogprob, fwdlattice)
last_column = fwdlattice[-1]
assert np.size(last_column) == n_state_combinations
score = logsumexp(last_column)
return score
def _do_forward_pass(self, framelogprob):
n_observations = framelogprob.shape[0]
state_combinations = [tuple(x) for x in list(itertools.product(np.arange(self.n_states), repeat=self.n_chains))]
fwdlattice = np.zeros((n_observations, self.n_states ** self.n_chains))
fhmmc._forward(n_observations, self.n_chains, self.n_states, state_combinations, self.log_startprob,
self.log_transmat, framelogprob, fwdlattice)
return logsumexp(fwdlattice[-1]), fwdlattice
def _do_forward_pass(self, framelogprob):
n_observations = framelogprob.shape[0]
state_combinations = [
tuple(x) for x in list(
itertools.product(np.arange(self.n_states),
repeat=self.n_chains))
]
fwdlattice = np.zeros((n_observations, self.n_states**self.n_chains))
fhmmc._forward(n_observations, self.n_chains, self.n_states,
state_combinations, self.log_startprob,
self.log_transmat, framelogprob, fwdlattice)
return logsumexp(fwdlattice[-1]), fwdlattice
def _exact_loglikelihood(self, ob):
log_transmat = np.zeros((self.n_chains, self.n_states, self.n_states))
log_startprob = np.zeros((self.n_chains, self.n_states))
for idx, chain in enumerate(self.chains_):
log_transmat[idx] = chain._log_transmat
log_startprob[idx] = chain._log_startprob
n_state_combinations = self.n_states**self.n_chains
state_combinations = [
tuple(x) for x in list(
itertools.product(np.arange(self.n_states),
repeat=self.n_chains))
]
n_observations = ob.shape[0]
n_features = ob.shape[1]
fwdlattice = np.zeros((n_observations, n_state_combinations))
# Calculate means and covariances for all state combinations and calculate emission probabilities
weight = (1.0 / float(self.n_chains))
weight_squared = weight * weight
covars = np.zeros(
(n_state_combinations,
n_features)) # TODO: add support for all covariance types
means = np.zeros((n_state_combinations, n_features))
for idx, state_combination in enumerate(state_combinations):
for chain_idx, state in enumerate(state_combination):
chain = self.chains_[chain_idx]
covars[idx] += chain._covars_[state]
means[idx] += chain._means_[state]
covars[idx] *= weight_squared
means[idx] *= weight
framelogprob = log_multivariate_normal_density(
ob, means, covars, covariance_type='diag'
) # TODO: add support for all covariance types
# Run the forward algorithm
fhmmc._forward(n_observations, self.n_chains, self.n_states,
state_combinations, log_startprob, log_transmat,
framelogprob, fwdlattice)
last_column = fwdlattice[-1]
assert np.size(last_column) == n_state_combinations
score = logsumexp(last_column)
return score
def test_forward_with_hmmlearn(self):
r = np.random.randn
obs = [np.array([[-600 + r(), 100 + r()], [-300 + r(), 200 + r()], [0 + r(), 300 + r()]]) for _ in xrange(10)]
hmm = GaussianHMM(n_components=3)
hmm.fit(obs)
# Calculcate fwdlattice using hmmlearn algorithm
framelogprob = hmm._compute_log_likelihood(obs[0])
start = timeit.default_timer()
_, fwdlattice1 = hmm._do_forward_pass(framelogprob)
print('hmmlearn took %fs' % (timeit.default_timer() - start))
# Calculate fwdlattice using fhmm algorithm with #chains = 1. This should yield the exact same results
start = timeit.default_timer()
fwdlattice2 = np.zeros(fwdlattice1.shape)
fhmmc._forward(obs[0].shape[0], 1, hmm.n_components, [(x,) for x in xrange(hmm.n_components)],
hmm._log_startprob.reshape(1, 3), hmm._log_transmat.reshape(1, 3, 3), framelogprob, fwdlattice2)
print('fhmm took %fs' % (timeit.default_timer() - start))
self.assertTrue(np.allclose(fwdlattice1, fwdlattice2))
