def solve_multinest(self, create_analyzer=True):
"""
Run pymultinest.solve, saving chain and parameter values
"""
parameters = self.model.labels
n_params = len(parameters)
result = solve(LogLikelihood=self.loglike, Prior=self.priors, n_dims=n_params, outputfiles_basename=self.prefix, verbose=True)
print()
print('evidence: %(logZ).1f +- %(logZerr).1f' % result)
print()
print('parameter values:')
self.params = []
for name, col in zip(parameters, result['samples'].transpose()):
self.params.append(col.mean())
print('%15s : %.3f +- %.3f' % (name, col.mean(), col.std()))
# make marginal plots by running:
# $ python multinest_marginals.py chains/3-
# For that, we need to store the parameter names:
import json
if not os.path.exists(self.prefix):
os.mkdir(self.prefix)
else:
for i in os.listdir('.'):
shutil.rmtree(self.prefix)
os.mkdir(self.prefix)
with open(os.path.join(self.prefix, '%sparams.json' % self.prefix), 'w') as f:
json.dump(parameters, f, indent=2)
for i in os.listdir('.'):
if i.startswith(self.prefix):
if os.path.isfile(i):
shutil.move(i,self.prefix+"/")
if create_analyzer == True:
self.analyzer = Analyzer(n_params, outputfiles_basename=self.prefix)
def deviance_ic(self):
"""Estimate Deviance Information Criterion.
This function estimates the Deviance Information Criterion (DIC) for the
model simulated with Nested Sampling (NS). It does so by using the
posterior distribution estimates computed from the NS outputs.
The DIC formula is given by:
DIC = p_D + D_bar,
where p_D = D_bar - D(theta_bar), D_bar is the posterior average of
the deviance D(theta)= -2*ln(L(theta)) with L(theta) the likelihood
of parameter set theta, and theta_bar is posterior average parameter set.
Returns:
float: The DIC estimate.
"""
mn_data = Analyzer(len(self.sampled_parameters),
self._file_root,
verbose=False).get_data()
params = mn_data[:, 2:]
log_likelihoods = -0.5 * mn_data[:, 1]
prior_mass = mn_data[:, 0]
norm_weights = (prior_mass * np.exp(log_likelihoods)) / self.evidence
nw_mask = np.isnan(norm_weights)
if np.any(nw_mask):
return np.inf
D_of_theta = -2. * log_likelihoods
D_bar = np.average(D_of_theta, weights=norm_weights)
theta_bar = np.average(params, axis=0, weights=norm_weights)
D_of_theta_bar = -2. * self.loglikelihood(theta_bar)
p_D = D_bar - D_of_theta_bar
return p_D + D_bar
def max_loglikelihood(self):
mn_data = Analyzer(len(self.sampled_parameters),
self._file_root,
verbose=False).get_data()
log_ls = -0.5 * mn_data[:, 1]
ml = log_ls.max()
return ml
def best_fit_likelihood(self):
"""Parameter vector with the maximum likelihood.
Returns:
numpy.array: The parameter vector.
"""
mn_data = Analyzer(len(self.sampled_parameters),
self._file_root,
verbose=False).get_data()
log_ls = -0.5 * mn_data[:, 1]
midx = np.argmax(log_ls)
ml = mn_data[midx][2:]
return ml
plt.ylabel(r'$r \times \Delta\Sigma$')
color_cycle = ['black', 'blue', 'green', 'red', 'brown']
style_cycle = ['--', '-.', ':', '-..']
for i, (multinest_dir, inv_cov_file, cov_file, label) in enumerate(zip(args.multinest_dir, args.inv_cov, args.cov, args.label)):
print(f"reading {multinest_dir}")
line_color = color_cycle[np.mod(i, len(color_cycle))]
line_style = style_cycle[np.mod(i, len(style_cycle))]
## read samples
n_dims = 14
a = Analyzer(n_dims, outputfiles_basename=multinest_dir)
bestfit_params = a.get_best_fit()['parameters']
bestfit_lnL = a.get_best_fit()['log_likelihood']
ppd_ml = np.array( bestfit_params[n_dims:] )
ml_chisq = ppd_ml[-1]
ml_wp, ml_ds = np.split(np.array(ppd_ml[:-1]), 2)
multinest_equal_samples_all = a.get_equal_weighted_posterior()
multinest_equal_samples = multinest_equal_samples_all[:, :n_dims]
multinest_ppd = multinest_equal_samples_all[:, n_dims:-1]
## compute PPD statistics
# importing modules
from __future__ import absolute_import, unicode_literals, print_function
from pymultinest.solve import solve
from pymultinest.analyse import Analyzer
import os
import sys
import shutil
import plot
import matplotlib.pyplot as plt
try:
os.mkdir('chains')
except OSError:
pass
do_pretty = True
n_params = 8
prefix = 'multi_fg_1-'
an = Analyzer(n_params, outputfiles_basename=prefix)
for i in range(n_params):
plt.figure()
marge = plot.PlotMarginalModes(an)
if do_pretty == True:
out = marge.plot_modes_marginal(i, with_ellipses=False)
plt.savefig("parameter%i_marginal.png" % i)
else:
out = marge.plot_conditional(i, with_ellipses=False)
plt.savefig("posterior%i_marginal.png" % i)
minimizer.set_cube(min_values, max_values)
minimizer.set_chi2_0()
# Make sure the directory with detailed results exists:
dir_out = "chains/"
if not os.path.exists(dir_out):
os.mkdir(dir_out)
file_prefix = __file__.split(".py")[0]
# Run MultiNest:
run_kwargs = {
'LogLikelihood': minimizer.ln_likelihood,
'Prior': minimizer.transform_cube,
'n_dims': len(parameters_to_fit),
'outputfiles_basename': os.path.join(dir_out, file_prefix+"_"),
'resume': False,
'importance_nested_sampling': False,
'multimodal': True}
result = solve(**run_kwargs)
# Analyze results - we print each mode separately and give log-evidence:
analyzer = Analyzer(n_params=run_kwargs['n_dims'],
outputfiles_basename=run_kwargs['outputfiles_basename'])
modes = analyzer.get_mode_stats()['modes']
for mode in modes:
print("\nMode {:} parameters:".format(mode['index']))
for (p, m, s) in zip(parameters_to_fit, mode['mean'], mode['sigma']):
print("{:6} = {:.4f} +- {:.4f}".format(p, m, s))
print("local log-evidence: {:.3f} +- {:.3f}".format(
mode["local log-evidence"], mode["local log-evidence error"]))
cpm_sources[i].pixel_time + 2450000.,
MCPM_options['mask_model_epochs'])
# MN fit:
dir_out = os.path.dirname(MN_args['outputfiles_basename'])
if not os.path.exists(dir_out):
os.mkdir(dir_out)
minimizer.set_MN_cube(mn_min, mn_max)
with redirect_stdout.stdout_redirect_2():
result = solve(LogLikelihood=minimizer.ln_like,
Prior=minimizer.transform_MN_cube,
**MN_args)
minimizer.close_file_all_models()
# Analyze output:
analyzer = Analyzer(n_params=MN_args['n_dims'],
outputfiles_basename=MN_args['outputfiles_basename'],
verbose=False)
stats = analyzer.get_stats()
msg = "Log-eveidence: {:.4f} +- {:.4f}"
log_evidence = stats['nested sampling global log-evidence']
log_evidence_err = stats['nested sampling global log-evidence error']
print(msg.format(log_evidence, log_evidence_err))
print('parameter values:')
for (name, v) in zip(parameters_to_fit, stats['marginals']):
median = v['median']
sigmas = [v['1sigma'][1] - median, median - v['1sigma'][0]]
print('{:7s} : {:.4f} {:.4f} {:.4f}'.format(name, median, *sigmas))
print('Best model:')
minimizer.print_min_chi2()
from scipy import stats
from pymultinest.analyse import Analyzer
llhs = []
bestfits = []
event = 2
for i in range(1000):
#if os.path.exists('/gpfs/scratch/pde3/retro/test_event%i/log/run_%04d.log'%(event,i)):
if os.path.exists('/gpfs/scratch/pde3/retro/log_cscd/run_%04d.log' % (i)):
try:
#a = Analyzer(8, outputfiles_basename="/gpfs/scratch/pde3/retro/test_event%i/out/tol0.1_evt%i-"%(event,i))
a = Analyzer(8,
outputfiles_basename=
"/gpfs/scratch/pde3/retro/out_cscd/tol0.1_evt%i-" %
(i))
bestfit_params = a.get_best_fit()
llhs.append(bestfit_params['log_likelihood'])
bestfits.append(bestfit_params['parameters'])
except IOError:
pass
llhs = np.array(llhs)
bestfits = np.array(bestfits)
names = [
'time', 'x', 'y', 'z', 'zenith', 'azimuth', 'energy', 'track_fraction'
]
units = ['ns', 'm', 'm', 'm', 'rad', 'rad', 'GeV', None]
#event 1:
