def est_covariance_mtx(self):
"""
Returns an estimate of the covariance of the current posterior model
distribution, given by the covariance of the current SMC approximation.
:rtype: :class:`numpy.ndarray`, shape
``(n_modelparams, n_modelparams)``.
:returns: An array containing the estimated covariance matrix.
"""
return particle_covariance_mtx(
self.particle_weights,
self.particle_locations)
def est_cluster_moments(self, cluster_opts=None):
# TODO: document
if cluster_opts is None:
cluster_opts = {}
for cluster_label, cluster_particles in particle_clusters(
self.particle_locations, self.particle_weights,
**cluster_opts):
w = self.particle_weights[cluster_particles]
l = self.particle_locations[cluster_particles]
yield (
cluster_label,
sum(w), # The zeroth moment is very useful here!
u.particle_meanfn(w, l, lambda x: x),
u.particle_covariance_mtx(w, l))
def est_cluster_moments(self, cluster_opts=None):
# TODO: document
if cluster_opts is None:
cluster_opts = {}
for cluster_label, cluster_particles in particle_clusters(
self.particle_locations, self.particle_weights,
**cluster_opts
):
w = self.particle_weights[cluster_particles]
l = self.particle_locations[cluster_particles]
yield (
cluster_label,
sum(w), # The zeroth moment is very useful here!
u.particle_meanfn(w, l, lambda x: x),
u.particle_covariance_mtx(w, l)
)
def est_cluster_covs(self, cluster_opts=None):
# TODO: document
cluster_moments = np.array(
list(self.est_cluster_moments(cluster_opts=cluster_opts)),
dtype=[
('label', 'int'),
('weight', 'float64'),
('mean', '{}float64'.format(self.n_rvs)),
('cov', '{0},{0}float64'.format(self.n_rvs)),
])
ws = cluster_moments['weight'][:, np.newaxis, np.newaxis]
within_cluster_var = np.sum(ws * cluster_moments['cov'], axis=0)
between_cluster_var = u.particle_covariance_mtx(
# Treat the cluster means as a new very small particle cloud.
cluster_moments['weight'], cluster_moments['mean']
)
total_var = within_cluster_var + between_cluster_var
return within_cluster_var, between_cluster_var, total_var
def est_cluster_covs(self, cluster_opts=None):
# TODO: document
cluster_moments = np.array(
list(self.est_cluster_moments(cluster_opts=cluster_opts)),
dtype=[
('label', 'int'),
('weight', 'float64'),
('mean', '{}float64'.format(self.n_rvs)),
('cov', '{0},{0}float64'.format(self.n_rvs)),
])
ws = cluster_moments['weight'][:, np.newaxis, np.newaxis]
within_cluster_var = np.sum(ws * cluster_moments['cov'], axis=0)
between_cluster_var = u.particle_covariance_mtx(
# Treat the cluster means as a new very small particle cloud.
cluster_moments['weight'],
cluster_moments['mean'])
total_var = within_cluster_var + between_cluster_var
return within_cluster_var, between_cluster_var, total_var
def est_covariance_mtx(self, corr=False):
"""
Returns the full-rank covariance matrix of the current particle
distribution.
:param bool corr: If `True`, the covariance matrix is normalized
by the outer product of the square root diagonal of the covariance matrix,
i.e. the correlation matrix is returned instead.
:rtype: :class:`numpy.ndarray`, shape
``(n_modelparams, n_modelparams)``.
:returns: An array containing the estimated covariance matrix.
"""
cov = u.particle_covariance_mtx(self.particle_weights,
self.particle_locations)
if corr:
dstd = np.sqrt(np.diag(cov))
cov /= (np.outer(dstd, dstd))
return cov
def est_covariance_mtx(self, corr=False):
"""
Returns the full-rank covariance matrix of the current particle
distribution.
:param bool corr: If `True`, the covariance matrix is normalized
by the outer product of the square root diagonal of the covariance matrix,
i.e. the correlation matrix is returned instead.
:rtype: :class:`numpy.ndarray`, shape
``(n_modelparams, n_modelparams)``.
:returns: An array containing the estimated covariance matrix.
"""
cov = u.particle_covariance_mtx(
self.particle_weights,
self.particle_locations)
if corr:
dstd = np.sqrt(np.diag(cov))
cov /= (np.outer(dstd, dstd))
return cov
def calc_covariance_matrix(self):
'''
Calculates the covariance matrix
'''
return utils.particle_covariance_mtx(self.weights, self.points)