def group_sparse_covariance(emp_covs, n_samples, alpha, max_iter, tol):
precisions = _group_sparse_covariance(
emp_covs, n_samples, alpha, max_iter=max_iter, tol=tol,
verbose=1, debug=False)
return (dict(n_samples=n_samples, alpha=alpha, max_iter=max_iter, tol=tol,
precisions=precisions))
def save_group_sparse_covariance(emp_covs, n_samples, alpha, max_iter, tol,
cache_dir, num=0, random_init=True,
debug=False):
if random_init:
rand_gen = np.random.RandomState(
int(int(1000000 * time.time()) % 100000000))
precisions_init = np.empty(emp_covs.shape)
for k in xrange(emp_covs.shape[-1]):
precisions_init[..., k] = random_spd(emp_covs.shape[0],
rand_gen=rand_gen)
else:
precisions_init = None
probe = ScoreProbe()
precisions = _group_sparse_covariance(
emp_covs, n_samples, alpha, max_iter=max_iter, tol=tol,
verbose=1, debug=debug, probe_function=probe,
precisions_init=precisions_init)
output_fname = os.path.join(cache_dir,
"precisions_{num:d}.pickle".format(num=num))
pickle.dump(dict(n_samples=n_samples, alpha=alpha, max_iter=max_iter,
tol=tol, objective=probe.objective,
log_lik=probe.log_lik, wall_clock=probe.wall_clock,
precisions=precisions, precisions_init=precisions_init),
open(output_fname, "wb"))
def benchmark(parameters, output_d="_convergence"):
_, _, gt = create_signals(parameters, output_dir=output_d)
emp_covs, n_samples = empirical_covariances(gt["signals"])
print("alpha_max: %.3e, %.3e" % compute_alpha_max(emp_covs, n_samples))
sp = ScoreProbe(duality_gap=True)
_group_sparse_covariance(
emp_covs, n_samples, alpha=parameters["alpha"], tol=parameters["tol"],
max_iter=parameters["max_iter"], probe_function=sp, verbose=1)
return {"log_lik": np.asarray(sp.log_lik),
"objective": np.asarray(sp.objective),
"precisions": np.asarray(sp.precisions),
"duality_gap": np.asarray(sp.duality_gap),
"time": np.asarray(sp.wall_clock)}, gt
