def _estimate(self, dtrajs):
"""
Return
------
hmsm : :class:`EstimatedHMSM <pyemma.msm.estimators.hmsm_estimated.EstimatedHMSM>`
Estimated Hidden Markov state model
"""
# ensure right format
dtrajs = ensure_dtraj_list(dtrajs)
# if no initial MSM is given, estimate it now
if self.init_hmsm is None:
# estimate with store_data=True, because we need an EstimatedHMSM
hmsm_estimator = _MaximumLikelihoodHMSM(
lag=self.lag,
stride=self.stride,
nstates=self.nstates,
reversible=self.reversible,
connectivity=self.connectivity,
observe_active=self.observe_active,
dt_traj=self.dt_traj)
init_hmsm = hmsm_estimator.estimate(
dtrajs) # estimate with lagged trajectories
else:
# check input
assert isinstance(
self.init_hmsm,
_EstimatedHMSM), 'hmsm must be of type EstimatedHMSM'
init_hmsm = self.init_hmsm
self.nstates = init_hmsm.nstates
self.reversible = init_hmsm.is_reversible
# here we blow up the output matrix (if needed) to the FULL state space because we want to use dtrajs in the
# Bayesian HMM sampler
if self.observe_active:
import msmtools.estimation as msmest
nstates_full = msmest.number_of_states(dtrajs)
# pobs = _np.zeros((init_hmsm.nstates, nstates_full)) # currently unused because that produces zero cols
eps = 0.01 / nstates_full # default output probability, in order to avoid zero columns
# full state space output matrix. make sure there are no zero columns
pobs = eps * _np.ones(
(self.nstates, nstates_full), dtype=_np.float64)
# fill active states
pobs[:, init_hmsm.observable_set] = _np.maximum(
eps, init_hmsm.observation_probabilities)
# renormalize B to make it row-stochastic
pobs /= pobs.sum(axis=1)[:, None]
else:
pobs = init_hmsm.observation_probabilities
# HMM sampler
if self.show_progress:
self._progress_register(self.nsamples,
description='Sampling HMSMs',
stage=0)
def call_back():
self._progress_update(1, stage=0)
else:
call_back = None
from bhmm import discrete_hmm, bayesian_hmm
hmm_mle = discrete_hmm(init_hmsm.transition_matrix,
pobs,
stationary=True,
reversible=self.reversible)
# define prior
if self.prior == 'sparse':
self.prior_count_matrix = _np.zeros((self.nstates, self.nstates),
dtype=_np.float64)
elif self.prior == 'uniform':
self.prior_count_matrix = _np.ones((self.nstates, self.nstates),
dtype=_np.float64)
elif self.prior == 'mixed':
# C0 = _np.dot(_np.diag(init_hmsm.stationary_distribution), init_hmsm.transition_matrix)
P0 = init_hmsm.transition_matrix
P0_offdiag = P0 - _np.diag(_np.diag(P0))
scaling_factor = 1.0 / _np.sum(P0_offdiag, axis=1)
self.prior_count_matrix = P0 * scaling_factor[:, None]
else:
raise ValueError('Unknown prior mode: ' + self.prior)
sampled_hmm = bayesian_hmm(
init_hmsm.discrete_trajectories_lagged,
hmm_mle,
nsample=self.nsamples,
transition_matrix_prior=self.prior_count_matrix,
call_back=call_back)
if self.show_progress:
self._progress_force_finish(stage=0)
# Samples
sample_Ps = [
sampled_hmm.sampled_hmms[i].transition_matrix
for i in range(self.nsamples)
]
sample_pis = [
sampled_hmm.sampled_hmms[i].stationary_distribution
for i in range(self.nsamples)
]
sample_pobs = [
sampled_hmm.sampled_hmms[i].output_model.output_probabilities
for i in range(self.nsamples)
]
samples = []
for i in range(
self.nsamples): # restrict to observable set if necessary
Bobs = sample_pobs[i][:, init_hmsm.observable_set]
sample_pobs[i] = Bobs / Bobs.sum(axis=1)[:, None] # renormalize
samples.append(
_HMSM(sample_Ps[i],
sample_pobs[i],
pi=sample_pis[i],
dt_model=init_hmsm.dt_model))
# parametrize self
self._dtrajs_full = dtrajs
self._observable_set = init_hmsm._observable_set
self._dtrajs_obs = init_hmsm._dtrajs_obs
self.set_model_params(samples=samples,
P=init_hmsm.transition_matrix,
pobs=init_hmsm.observation_probabilities,
dt_model=init_hmsm.dt_model)
return self
def _estimate(self, dtrajs):
# ensure right format
dtrajs = ensure_dtraj_list(dtrajs)
if self.init_hmsm is None: # estimate using maximum-likelihood superclass
# memorize the observation state for bhmm and reset
# TODO: more elegant solution is to set Estimator params only temporarily in estimate(X, **kwargs)
default_connectivity = self.connectivity
default_mincount_connectivity = self.mincount_connectivity
default_observe_nonempty = self.observe_nonempty
self.connectivity = None
self.observe_nonempty = False
self.mincount_connectivity = 0
self.accuracy = 1e-2 # this is sufficient for an initial guess
super(BayesianHMSM, self)._estimate(dtrajs)
self.connectivity = default_connectivity
self.mincount_connectivity = default_mincount_connectivity
self.observe_nonempty = default_observe_nonempty
else: # if given another initialization, must copy its attributes
# TODO: this is too tedious - need to automatize parameter+result copying between estimators.
self.nstates = self.init_hmsm.nstates
self.reversible = self.init_hmsm.is_reversible
self.stationary = self.init_hmsm.stationary
# trajectories
self._dtrajs_full = self.init_hmsm._dtrajs_full
self._dtrajs_lagged = self.init_hmsm._dtrajs_lagged
self._observable_set = self.init_hmsm._observable_set
self._dtrajs_obs = self.init_hmsm._dtrajs_obs
# MLE estimation results
self.likelihoods = self.init_hmsm.likelihoods # Likelihood history
self.likelihood = self.init_hmsm.likelihood
self.hidden_state_probabilities = self.init_hmsm.hidden_state_probabilities # gamma variables
self.hidden_state_trajectories = self.init_hmsm.hidden_state_trajectories # Viterbi path
self.count_matrix = self.init_hmsm.count_matrix # hidden count matrix
self.initial_count = self.init_hmsm.initial_count # hidden init count
self.initial_distribution = self.init_hmsm.initial_distribution
self._active_set = self.init_hmsm._active_set
# update HMM Model
self.update_model_params(
P=self.init_hmsm.transition_matrix,
pobs=self.init_hmsm.observation_probabilities,
dt_model=TimeUnit(self.dt_traj).get_scaled(self.lag))
# check if we have a valid initial model
import msmtools.estimation as msmest
if self.reversible and not msmest.is_connected(self.count_matrix):
raise NotImplementedError(
'Encountered disconnected count matrix:\n ' +
str(self.count_matrix) +
'with reversible Bayesian HMM sampler using lag=' +
str(self.lag) + ' and stride=' + str(self.stride) +
'. Consider using shorter lag, ' +
'or shorter stride (to use more of the data), ' +
'or using a lower value for mincount_connectivity.')
# here we blow up the output matrix (if needed) to the FULL state space because we want to use dtrajs in the
# Bayesian HMM sampler. This is just an initialization.
import msmtools.estimation as msmest
nstates_full = msmest.number_of_states(dtrajs)
if self.nstates_obs < nstates_full:
eps = 0.01 / nstates_full # default output probability, in order to avoid zero columns
# full state space output matrix. make sure there are no zero columns
B_init = eps * _np.ones(
(self.nstates, nstates_full), dtype=_np.float64)
# fill active states
B_init[:, self.observable_set] = _np.maximum(
eps, self.observation_probabilities)
# renormalize B to make it row-stochastic
B_init /= B_init.sum(axis=1)[:, None]
else:
B_init = self.observation_probabilities
# HMM sampler
if self.show_progress:
self._progress_register(self.nsamples,
description='Sampling HMSMs',
stage=0)
def call_back():
self._progress_update(1, stage=0)
else:
call_back = None
from bhmm import discrete_hmm, bayesian_hmm
hmm_mle = discrete_hmm(self.initial_distribution,
self.transition_matrix, B_init)
sampled_hmm = bayesian_hmm(
self.discrete_trajectories_lagged,
hmm_mle,
nsample=self.nsamples,
reversible=self.reversible,
stationary=self.stationary,
p0_prior=self.p0_prior,
transition_matrix_prior=self.transition_matrix_prior,
store_hidden=self.store_hidden,
call_back=call_back)
if self.show_progress:
self._progress_force_finish(stage=0)
# Samples
sample_Ps = [
sampled_hmm.sampled_hmms[i].transition_matrix
for i in range(self.nsamples)
]
sample_pis = [
sampled_hmm.sampled_hmms[i].stationary_distribution
for i in range(self.nsamples)
]
sample_pobs = [
sampled_hmm.sampled_hmms[i].output_model.output_probabilities
for i in range(self.nsamples)
]
samples = []
for i in range(
self.nsamples): # restrict to observable set if necessary
Bobs = sample_pobs[i][:, self.observable_set]
sample_pobs[i] = Bobs / Bobs.sum(axis=1)[:, None] # renormalize
samples.append(
_HMSM(sample_Ps[i],
sample_pobs[i],
pi=sample_pis[i],
dt_model=self.dt_model))
# store results
self.sampled_trajs = [
sampled_hmm.sampled_hmms[i].hidden_state_trajectories
for i in range(self.nsamples)
]
self.update_model_params(samples=samples)
# deal with connectivity
states_subset = None
if self.connectivity == 'largest':
states_subset = 'largest-strong'
elif self.connectivity == 'populous':
states_subset = 'populous-strong'
# OBSERVATION SET
if self.observe_nonempty:
observe_subset = 'nonempty'
else:
observe_subset = None
# return submodel (will return self if all None)
return self.submodel(states=states_subset,
obs=observe_subset,
mincount_connectivity=self.mincount_connectivity)
def _estimate(self, dtrajs):
# ensure right format
dtrajs = ensure_dtraj_list(dtrajs)
if self.init_hmsm is None: # estimate using maximum-likelihood superclass
# memorize the observation state for bhmm and reset
# TODO: more elegant solution is to set Estimator params only temporarily in estimate(X, **kwargs)
default_connectivity = self.connectivity
default_mincount_connectivity = self.mincount_connectivity
default_observe_nonempty = self.observe_nonempty
self.connectivity = None
self.observe_nonempty = False
self.mincount_connectivity = 0
self.accuracy = 1e-2 # this is sufficient for an initial guess
super(BayesianHMSM, self)._estimate(dtrajs)
self.connectivity = default_connectivity
self.mincount_connectivity = default_mincount_connectivity
self.observe_nonempty = default_observe_nonempty
else: # if given another initialization, must copy its attributes
copy_attributes = ['_nstates', '_reversible', '_pi', '_observable_set', 'likelihoods', 'likelihood',
'hidden_state_probabilities', 'hidden_state_trajectories', 'count_matrix',
'initial_count', 'initial_distribution', '_active_set']
check_user_choices = ['lag', '_nstates']
# check if nstates and lag are compatible
for attr in check_user_choices:
if not getattr(self, attr) == getattr(self.init_hmsm, attr):
raise UserWarning('BayesianHMSM cannot be initialized with init_hmsm with '
'incompatible lag or nstates.')
if (len(dtrajs) != len(self.init_hmsm.dtrajs_full) or
not all((_np.array_equal(d1, d2) for d1, d2 in zip(dtrajs, self.init_hmsm.dtrajs_full)))):
raise NotImplementedError('Bayesian HMM estimation with init_hmsm is currently only implemented ' +
'if applied to the same data.')
# TODO: implement more elegant solution to copy-pasting effective stride evaluation from ML HMM.
# EVALUATE STRIDE
if self.stride == 'effective':
# by default use lag as stride (=lag sampling), because we currently have no better theory for deciding
# how many uncorrelated counts we can make
self.stride = self.lag
# get a quick estimate from the spectral radius of the nonreversible
from pyemma.msm import estimate_markov_model
msm_nr = estimate_markov_model(dtrajs, lag=self.lag, reversible=False, sparse=False,
connectivity='largest', dt_traj=self.timestep_traj)
# if we have more than nstates timescales in our MSM, we use the next (neglected) timescale as an
# estimate of the decorrelation time
if msm_nr.nstates > self.nstates:
corrtime = max(1, msm_nr.timescales()[self.nstates - 1])
# use the smaller of these two pessimistic estimates
self.stride = int(min(self.lag, 2 * corrtime))
# if stride is different to init_hmsm, check if microstates in lagged-strided trajs are compatible
if self.stride != self.init_hmsm.stride:
dtrajs_lagged_strided = _lag_observations(dtrajs, self.lag, stride=self.stride)
_nstates_obs = _number_of_states(dtrajs_lagged_strided, only_used=True)
_nstates_obs_full = _number_of_states(dtrajs)
if _np.setxor1d(_np.concatenate(dtrajs_lagged_strided),
_np.concatenate(self.init_hmsm._dtrajs_lagged)).size != 0:
raise UserWarning('Choice of stride has excluded a different set of microstates than in ' +
'init_hmsm. Set of observed microstates in time-lagged strided trajectories ' +
'must match to the one used for init_hmsm estimation.')
self._dtrajs_full = dtrajs
self._dtrajs_lagged = dtrajs_lagged_strided
self._nstates_obs_full = _nstates_obs_full
self._nstates_obs = _nstates_obs
self._observable_set = _np.arange(self._nstates_obs)
self._dtrajs_obs = dtrajs
else:
copy_attributes += ['_dtrajs_full', '_dtrajs_lagged', '_nstates_obs_full',
'_nstates_obs', '_observable_set', '_dtrajs_obs']
# update self with estimates from init_hmsm
self.__dict__.update(
{k: i for k, i in self.init_hmsm.__dict__.items() if k in copy_attributes})
# as mentioned in the docstring, take init_hmsm observed set observation probabilities
self.observe_nonempty = False
# update HMM Model
self.update_model_params(P=self.init_hmsm.transition_matrix, pobs=self.init_hmsm.observation_probabilities,
dt_model=TimeUnit(self.dt_traj).get_scaled(self.lag))
# check if we have a valid initial model
import msmtools.estimation as msmest
if self.reversible and not msmest.is_connected(self.count_matrix):
raise NotImplementedError('Encountered disconnected count matrix:\n ' + str(self.count_matrix)
+ 'with reversible Bayesian HMM sampler using lag=' + str(self.lag)
+ ' and stride=' + str(self.stride) + '. Consider using shorter lag, '
+ 'or shorter stride (to use more of the data), '
+ 'or using a lower value for mincount_connectivity.')
# here we blow up the output matrix (if needed) to the FULL state space because we want to use dtrajs in the
# Bayesian HMM sampler. This is just an initialization.
nstates_full = msmest.number_of_states(dtrajs)
if self.nstates_obs < nstates_full:
eps = 0.01 / nstates_full # default output probability, in order to avoid zero columns
# full state space output matrix. make sure there are no zero columns
B_init = eps * _np.ones((self.nstates, nstates_full), dtype=_np.float64)
# fill active states
B_init[:, self.observable_set] = _np.maximum(eps, self.observation_probabilities)
# renormalize B to make it row-stochastic
B_init /= B_init.sum(axis=1)[:, None]
else:
B_init = self.observation_probabilities
# HMM sampler
if self.show_progress:
self._progress_register(self.nsamples, description='Sampling HMSMs', stage=0)
def call_back():
self._progress_update(1, stage=0)
else:
call_back = None
from bhmm import discrete_hmm, bayesian_hmm
if self.init_hmsm is not None:
hmm_mle = self.init_hmsm.hmm
else:
hmm_mle = discrete_hmm(self.initial_distribution, self.transition_matrix, B_init)
sampled_hmm = bayesian_hmm(self.discrete_trajectories_lagged, hmm_mle, nsample=self.nsamples,
reversible=self.reversible, stationary=self.stationary,
p0_prior=self.p0_prior, transition_matrix_prior=self.transition_matrix_prior,
store_hidden=self.store_hidden, call_back=call_back)
if self.show_progress:
self._progress_force_finish(stage=0)
# Samples
sample_inp = [(m.transition_matrix, m.stationary_distribution, m.output_probabilities)
for m in sampled_hmm.sampled_hmms]
samples = []
for P, pi, pobs in sample_inp: # restrict to observable set if necessary
Bobs = pobs[:, self.observable_set]
pobs = Bobs / Bobs.sum(axis=1)[:, None] # renormalize
samples.append(_HMSM(P, pobs, pi=pi, dt_model=self.dt_model))
# store results
self.sampled_trajs = [sampled_hmm.sampled_hmms[i].hidden_state_trajectories for i in range(self.nsamples)]
self.update_model_params(samples=samples)
# deal with connectivity
states_subset = None
if self.connectivity == 'largest':
states_subset = 'largest-strong'
elif self.connectivity == 'populous':
states_subset = 'populous-strong'
# OBSERVATION SET
if self.observe_nonempty:
observe_subset = 'nonempty'
else:
observe_subset = None
# return submodel (will return self if all None)
return self.submodel(states=states_subset, obs=observe_subset,
mincount_connectivity=self.mincount_connectivity)