def likelihood(self, outcomes, modelparams, expparams):
# FIXME: at present, will proceed until ALL model experiment pairs
# are below error tol.
# Should disable one-by-one, but that's tricky.
super(ALEApproximateModel, self).likelihood(outcomes, modelparams,
expparams)
simulator = self.underlying_model
# We will use the fact we have assumed a two-outcome model to make the
# problem easier. As such, we will rely on the static method
# FiniteOutcomeModel.pr0_to_likelihood_array.
# Start off with min_samp samples.
n = np.zeros((modelparams.shape[0], expparams.shape[0]))
for N in count(start=self._min_samp, step=self._samp_step):
sim_data = simulator.simulate_experiment(modelparams,
expparams,
repeat=self._samp_step)
n += np.sum(sim_data,
axis=0) # Sum over the outcomes axis to find the
# number of 1s.
error_est_p1 = binom_est_error(
binom_est_p(n, N, self._adapt_hedge), N, self._adapt_hedge)
if np.all(error_est_p1 < self._error_tol): break
return FiniteOutcomeModel.pr0_to_likelihood_array(
outcomes, 1 - binom_est_p(n, N, self._est_hedge))
def likelihood(self, outcomes, modelparams, expparams):
# By calling the superclass implementation, we can consolidate
# call counting there.
# Get the original, undisturbed likelihoods.
super(PoisonedModel, self).likelihood(outcomes, modelparams, expparams)
L = self.underlying_model.likelihood(
outcomes, modelparams, expparams)
# Now get the random variates from a standard normal [N(0, 1)]
# distribution; we'll rescale them soon.
epsilon = np.random.normal(size=L.shape)
# If ALE, rescale by a constant tolerance.
if self._mode == PoisonModes.ALE:
epsilon *= self._tol
# Otherwise, rescale by the estimated error in the binomial estimator.
elif self._mode == PoisonModes.MLE:
epsilon *= binom_est_error(p=L, N=self._n_samples, hedge=self._hedge)
# Now we truncate and return.
np.clip(L + epsilon, 0, 1, out=L)
return L
def likelihood(self, outcomes, modelparams, expparams):
# FIXME: at present, will proceed until ALL model experiment pairs
# are below error tol.
# Should disable one-by-one, but that's tricky.
super(ALEApproximateModel, self).likelihood(outcomes, modelparams, expparams)
simulator = self.underlying_model
# We will use the fact we have assumed a two-outcome model to make the
# problem easier. As such, we will rely on the static method
# FiniteOutcomeModel.pr0_to_likelihood_array.
# Start off with min_samp samples.
n = np.zeros((modelparams.shape[0], expparams.shape[0]))
for N in count(start=self._min_samp, step=self._samp_step):
sim_data = simulator.simulate_experiment(
modelparams, expparams, repeat=self._samp_step
)
n += np.sum(sim_data, axis=0) # Sum over the outcomes axis to find the
# number of 1s.
error_est_p1 = binom_est_error(binom_est_p(n, N, self._adapt_hedge), N, self._adapt_hedge)
if np.all(error_est_p1 < self._error_tol): break
return FiniteOutcomeModel.pr0_to_likelihood_array(outcomes, 1 - binom_est_p(n, N, self._est_hedge))
def likelihood(self, outcomes, modelparams, expparams):
# By calling the superclass implementation, we can consolidate
# call counting there.
# Get the original, undisturbed likelihoods.
super(PoisonedModel, self).likelihood(outcomes, modelparams, expparams)
L = self.underlying_model.likelihood(outcomes, modelparams, expparams)
# Now get the random variates from a standard normal [N(0, 1)]
# distribution; we'll rescale them soon.
epsilon = np.random.normal(size=L.shape)
# If ALE, rescale by a constant tolerance.
if self._mode == PoisonModes.ALE:
epsilon *= self._tol
# Otherwise, rescale by the estimated error in the binomial estimator.
elif self._mode == PoisonModes.MLE:
epsilon *= binom_est_error(p=L,
N=self._n_samples,
hedge=self._hedge)
# Now we truncate and return.
np.clip(L + epsilon, 0, 1, out=L)
return L
