def initialize_model():
### @export 'load model'
dm = dismod3.load_disease_model(19807)
### @export 'initialize model data'
dm.params['global_priors']['level_bounds']['excess_mortality'] = dict(
lower=.1, upper=100.)
dm.params['global_priors']['increasing']['excess_mortality'] = dict(
age_start=0, age_end=0)
dm.params['global_priors']['level_bounds']['relative_risk'] = dict(
lower=0., upper=10000.)
for cv in dm.params['covariates']['Study_level']:
dm.params['covariates']['Study_level'][cv]['rate']['value'] = 0
for cv in dm.params['covariates']['Country_level']:
dm.params['covariates']['Country_level'][cv]['rate']['value'] = 0
level = .001
dm.params['sex_effect_prevalence'] = dict(mean=1,
upper_ci=pl.exp(level * 1.96),
lower_ci=pl.exp(-level * 1.96))
dm.params['time_effect_prevalence'] = dict(mean=1,
upper_ci=pl.exp(level * 1.96),
lower_ci=pl.exp(-level * 1.96))
dm.params['region_effect_prevalence'] = dict(std=level)
dm.clear_fit()
dm.clear_empirical_prior()
dismod3.neg_binom_model.covariate_hash = {}
return dm
def download_model(id):
""" Copy model from web to j drive
Parameters
----------
id : int
The model id number to copy
"""
dir = dismod3.settings.JOB_WORKING_DIR % id # TODO: refactor into a function
try:
model = dismod3.data.ModelData.load(dir)
print 'model already on j drive in %s' % dir
except (IOError, AssertionError):
print 'downloading disease model'
dm = dismod3.load_disease_model(id)
import simplejson as json
try:
model = dismod3.data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
except Exception as e:
print e
print 'attempting to use old covariate format'
import old_cov_data
model = old_cov_data.from_gbd_jsons(json.loads(dm.to_json()))
model.save(dir)
print 'loaded data from json, saved in new format for next time in %s' % dir
def initialize_model():
### @export 'load model'
dm = dismod3.load_disease_model(16391)
### @export 'set expert priors'
dm.set_param_age_mesh(pl.arange(0,101,10))
dm.params['global_priors']['smoothness']['prevalence']['amount'] = 'Moderately'
dm.params['global_priors']['heterogeneity']['prevalence'] = 'Slightly'
dm.params['global_priors']['level_value']['prevalence'] = dict(value=0., age_before=0, age_after=100)
dm.params['global_priors']['level_bounds']['prevalence'] = dict(lower=0., upper =.1)
dm.params['global_priors']['increasing']['prevalence'] = dict(age_start=0, age_end=0)
dm.params['global_priors']['decreasing']['prevalence'] = dict(age_start=100, age_end=100)
dm.params['covariates']['Study_level']['bias']['rate']['value'] = 1
for cv in dm.params['covariates']['Country_level']:
dm.params['covariates']['Country_level'][cv]['rate']['value'] = 0
### @export 'initialize model data'
dm.data = [d for d in dm.data if dm.relevant_to(d, 'prevalence', region, 'all', 'all')]
# fit model
dm.clear_fit()
dm.clear_empirical_prior()
dismod3.neg_binom_model.covariate_hash = {}
return dm
def initialize_model():
### @export 'load model'
dm = dismod3.load_disease_model(16391)
### @export 'set expert priors'
dm.set_param_age_mesh(pl.arange(0, 101, 10))
dm.params['global_priors']['smoothness']['prevalence'][
'amount'] = 'Slightly'
dm.params['global_priors']['heterogeneity']['prevalence'] = 'Slightly'
dm.params['global_priors']['level_value']['prevalence'] = dict(
value=0., age_before=0, age_after=100)
dm.params['global_priors']['level_bounds']['prevalence'] = dict(lower=0.,
upper=.1)
dm.params['global_priors']['increasing']['prevalence'] = dict(age_start=0,
age_end=0)
dm.params['global_priors']['decreasing']['prevalence'] = dict(
age_start=100, age_end=100)
dm.params['covariates']['Study_level']['bias']['rate']['value'] = 1
for cv in dm.params['covariates']['Country_level']:
dm.params['covariates']['Country_level'][cv]['rate']['value'] = 0
### @export 'initialize model data'
dm.data = [
d for d in dm.data
if dm.relevant_to(d, 'prevalence', region, year, 'all')
]
# fit model
dm.clear_fit()
dm.clear_empirical_prior()
dismod3.neg_binom_model.covariate_hash = {}
return dm
def validate_prior_similarity():
#dm = dismod3.load_disease_model(20945)
#dm.model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
#t = 'i'
#area, sex, year = 'europe_eastern', 'male', 2005
dm = dismod3.load_disease_model(20928)
dm.model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
t = 'p'
area, sex, year = 'sub-saharan_africa_central', 'male', 2005
# select data that is about areas in this region, recent years, and sex of male or total only
model = dm.model
subtree = nx.traversal.bfs_tree(model.hierarchy, area)
relevant_rows = [i for i, r in model.input_data.T.iteritems() \
if (r['area'] in subtree or r['area'] == 'all')\
and ((year == 2005 and r['year_end'] >= 1997) or r['year_start'] <= 1997) \
and r['sex'] in [sex, 'total']]
model.input_data = model.input_data.ix[relevant_rows]
# replace area 'all' with area
model.input_data['area'][model.input_data['area'] == 'all'] = area
for het in 'Slightly Moderately Very'.split():
dm.model.parameters[t]['parameter_age_mesh'] = [0, 15, 20, 25, 35, 45, 55, 65, 75, 100]
dm.model.parameters[t]['heterogeneity'] = het
setup_regional_model(dm, area, sex, year)
dm.vars = {}
dm.vars[t] = data_model.data_model(t, dm.model, t,
root_area=area, root_sex=sex, root_year=year,
mu_age=None,
mu_age_parent=dm.emp_priors[t, 'mu'],
sigma_age_parent=dm.emp_priors[t, 'sigma'],
rate_type=(t == 'rr') and 'log_normal' or 'neg_binom')
fit_model.fit_data_model(dm.vars[t], iter=10050, burn=5000, thin=50, tune_interval=100)
#2graphics.plot_one_effects(dm.vars[t], t, dm.model.hierarchy)
#pl.title(het)
graphics.plot_convergence_diag(dm.vars[t])
pl.title(het)
#graphics.plot_one_ppc(dm.vars[t], t)
#pl.title(het)
graphics.plot_one_type(dm.model, dm.vars[t], dm.emp_priors, t)
pl.title(het)
pl.show()
return dm
def initialize_model():
### @export 'load model'
dm = dismod3.load_disease_model(19271)
### @export 'initialize model data'
dm.data = [d for d in dm.data if dm.relevant_to(d, type, region, year, sex)]
for d in dm.data:
d['standard_error'] = float(d['sd_1enadj'] or d['parameter_value_old'])/10000. / pl.sqrt(d['effective_sample_size'])
d.pop('effective_sample_size')
# fit model
dm.clear_fit()
dm.clear_empirical_prior()
dismod3.neg_binom_model.covariate_hash = {}
return dm
def fit_emp_prior(id, param_type):
""" Fit empirical prior of specified type for specified model
Parameters
----------
id : int
The model id number for the job to fit
param_type : str, one of incidence, prevalence, remission, excess-mortality
The disease parameter to generate empirical priors for
Example
-------
>>> import fit_emp_prior
>>> fit_emp_prior.fit_emp_prior(2552, 'incidence')
"""
#dismod3.log_job_status(id, 'empirical_priors', param_type, 'Running')
# load disease model
dm = dismod3.load_disease_model(id)
#dm.data = [] # remove all data to speed up computation, for test
import dismod3.neg_binom_model as model
dir = dismod3.settings.JOB_WORKING_DIR % id
model.fit_emp_prior(dm, param_type, dbname='%s/empirical_priors/pickle/dm-%d-emp_prior-%s.pickle' % (dir, id, param_type))
# generate empirical prior plots
from pylab import subplot
for sex in dismod3.settings.gbd_sexes:
for year in dismod3.settings.gbd_years:
keys = dismod3.utils.gbd_keys(region_list=['all'], year_list=[year], sex_list=[sex], type_list=[param_type])
dismod3.tile_plot_disease_model(dm, keys, defaults={})
dm.savefig('dm-%d-emp_prior-%s-%s-%s.png' % (id, param_type, sex, year))
# TODO: put this in a separate script, which runs after all empirical priors are computed
for effect in ['alpha', 'beta', 'gamma', 'delta']:
dismod3.plotting.plot_empirical_prior_effects([dm], effect)
dm.savefig('dm-%d-emp-prior-%s-%s.png' % (id, param_type, effect))
# summarize fit quality graphically, as well as parameter posteriors
k0 = keys[0]
dm.vars = {k0: dm.vars} # hack to make posterior predictions plot
dismod3.plotting.plot_posterior_predicted_checks(dm, k0)
dm.savefig('dm-%d-emp-prior-check-%s.png' % (dm.id, param_type))
dm.vars = dm.vars[k0] # undo hack to make posterior predictions plot
# save results (do this last, because it removes things from the disease model that plotting function, etc, might need
dm.save('dm-%d-prior-%s.json' % (id, param_type))
dismod3.try_posting_disease_model(dm, ntries=5)
#dismod3.log_job_status(id, 'empirical_priors', param_type, 'Completed')
return dm
def find_fnrfx(model, disease, data_type, country, sex, year):
'''add fixed and random effects from GBD as priors to new model'''
# create dummy model to get appropriate Model.vars fields
dummy = load_new_model(disease, country, sex)
dummy.vars += dismod3.ism.age_specific_rate(dummy, data_type)
vars = dummy.vars[data_type]
# save random effects
try:
emp_re = pandas.read_csv(
'/home/j/Project/dismod/output/dm-%s/posterior/re-%s-%s+%s+%s.csv'
% (disease, data_type, geo_info(country, disease), sex, year),
index_col=0)
for col in emp_re.index:
model.parameters[data_type]['random_effects'][col] = dict(
dist='Constant',
mu=emp_re.ix[col, 'mu_coeff'],
sigma=emp_re.ix[col, 'sigma_coeff'])
except:
pass
# also save empirical prior on sigma_alpha, the dispersion of the random effects
dm = dismod3.load_disease_model(disease)
for n in vars['sigma_alpha']:
try:
dm_na = dm.get_empirical_prior(full_name[data_type])['new_alpha']
model.parameters[data_type]['random_effects'][n.__name__] = dict(
dist=dm_na[n.__name__]['dist'],
mu=dm_na[n.__name__]['mu'],
sigma=dm_na[n.__name__]['sigma'],
lower=dm_na[n.__name__]['lower'],
upper=dm_na[n.__name__]['upper'])
except:
model.parameters[data_type]['random_effects'][n.__name__] = dict(
dist='TruncatedNormal',
mu=.05,
sigma=.03**-2,
lower=0.01,
upper=0.5)
# save fixed effects
emp_fe = pandas.read_csv(
'/home/j/Project/dismod/output/dm-%s/posterior/fe-%s-%s+%s+%s.csv' %
(disease, data_type, geo_info(country, disease), sex, year),
index_col=0)
for n, col in zip(vars['beta'], vars['X'].columns):
model.parameters[data_type]['fixed_effects'][col] = dict(
dist='Constant',
mu=emp_fe.ix[col, 'mu_coeff'],
sigma=emp_fe.ix[col, 'sigma_coeff'])
def fetch_disease_model_if_necessary(id, dir_name):
try:
model = ModelData.load(dir_name)
print 'loaded data from new format from %s' % dir_name
except (IOError, AssertionError):
import os
os.makedirs(dir_name)
import dismod3.disease_json
dm = dismod3.load_disease_model(id)
import simplejson as json
model = ModelData.from_gbd_jsons(json.loads(dm.to_json()))
model.save(dir_name)
print 'loaded data from json, saved in new format for next time in %s' % dir_name
print 'model has %d rows of input data' % len(model.input_data.index)
return model
def fetch_disease_model_if_necessary(id, dir_name):
try:
model = ModelData.load(dir_name)
print 'loaded data from new format from %s' % dir_name
except (IOError, AssertionError):
import os
os.makedirs(dir_name)
import dismod3.disease_json
dm = dismod3.load_disease_model(id)
import simplejson as json
model = ModelData.from_gbd_jsons(json.loads(dm.to_json()))
model.save(dir_name)
print 'loaded data from json, saved in new format for next time in %s' % dir_name
print 'model has %d rows of input data' % len(model.input_data.index)
return model
def main():
import optparse
usage = 'usage: %prog [options] disease_model_id'
parser = optparse.OptionParser(usage)
parser.add_option('-s', '--sex', default='male',
help='only estimate given sex (valid settings ``male``, ``female``, ``all``)')
parser.add_option('-y', '--year', default='2005',
help='only estimate given year (valid settings ``1990``, ``2005``, ``2010``)')
parser.add_option('-r', '--region', default='australasia',
help='only estimate given GBD Region')
parser.add_option('-f', '--fast', default='False',
help='use MAP only')
parser.add_option('-i', '--inconsistent', default='False',
help='use inconsistent model for posteriors')
parser.add_option('-t', '--types', default='p i r',
help='with rate types to fit (only used if inconsistent=true)')
parser.add_option('-z', '--zerore', default='true',
help='enforce zero constraint on random effects')
parser.add_option('-o', '--onlyposterior', default='False',
help='skip empirical prior phase')
(options, args) = parser.parse_args()
if len(args) != 1:
parser.error('incorrect number of arguments')
try:
id = int(args[0])
except ValueError:
parser.error('disease_model_id must be an integer')
dm = dismod3.load_disease_model(id)
# set model id to passed-in id (should not be necessary)
dm.id = id
assert id == dm.id, 'model id should equal parameter id'
dm = fit_posterior(dm, options.region, options.sex, options.year,
fast_fit=options.fast.lower() == 'true',
inconsistent_fit=options.inconsistent.lower() == 'true',
params_to_fit=options.types.split(),
posteriors_only=(options.onlyposterior.lower()=='true'),
zero_re=options.zerore.lower() == 'true')
return dm
def fit_posterior(id, region, sex, year):
""" Fit posterior of specified region/sex/year for specified model
Parameters
----------
id : int
The model id number for the job to fit
region : str
From dismod3.settings.gbd_regions, but clean()-ed
sex : str, from dismod3.settings.gbd_sexes
year : str, from dismod3.settings.gbd_years
Example
-------
>>> import fit_posterior
>>> fit_posterior.fit_posterior(2552, 'asia_east', 'male', '2005')
"""
#print 'updating job status on server'
#dismod3.log_job_status(id, 'posterior', '%s--%s--%s' % (region, sex, year), 'Running')
dm = dismod3.load_disease_model(id)
#dm.data = [] # for testing, remove all data
keys = dismod3.utils.gbd_keys(region_list=[region], year_list=[year], sex_list=[sex], type_list=['prevalence'])
# fit the model
dir = dismod3.settings.JOB_WORKING_DIR % id
import dismod3.neg_binom_model as model
k0 = keys[0]
dm.vars = {}
dm.vars[k0] = model.setup(dm, k0, dm.data)
dm.mcmc = mc.MCMC(dm.vars)
dm.mcmc.sample(iter=50000,burn=25000,thin=50,verbose=1)
dm.map = mc.MAP(dm.vars)
dm.map.fit()
model.store_mcmc_fit(dm, k0, dm.vars[k0])
# update job status file
#print 'updating job status on server'
#dismod3.log_job_status(id, 'posterior',
# '%s--%s--%s' % (region, sex, year), 'Completed')
# save results (do this last, because it removes things from the disease model that plotting function, etc, might need
keys = dismod3.utils.gbd_keys(region_list=[region], year_list=[year], sex_list=[sex], type_list=['prevalence'])
dm.save('dm-%d-posterior-%s-%s-%s.json' % (id, region, sex, year), keys_to_save=keys)
return dm
def upload_fits(id):
""" Send results of cluster fits to dismod server
Parameters
----------
id : int
The model id number
Example
-------
>>> import fit_emp_prior
>>> fit_emp_prior.fit_emp_prior(2552, 'incidence')
>>> import upload_fits
>>> upload_fits.upload_fits(2552)
"""
# load disease model
dm = dismod3.load_disease_model(id) # this merges together results from all fits
dismod3.try_posting_disease_model(dm, ntries=5)
def find_fnrfx(model, disease, data_type, country, sex, year):
'''add fixed and random effects from GBD as priors to new model'''
# create dummy model to get appropriate Model.vars fields
dummy = load_new_model(disease, country, sex)
dummy.vars += dismod3.ism.age_specific_rate(dummy, data_type)
vars = dummy.vars[data_type]
# save random effects
try:
emp_re = pandas.read_csv('/home/j/Project/dismod/output/dm-%s/posterior/re-%s-%s+%s+%s.csv'%(disease, data_type, geo_info(country,disease), sex, year), index_col=0)
for col in emp_re.index:
model.parameters[data_type]['random_effects'][col] = dict(dist='Constant',
mu=emp_re.ix[col, 'mu_coeff'],
sigma=emp_re.ix[col, 'sigma_coeff'])
except:
pass
# also save empirical prior on sigma_alpha, the dispersion of the random effects
dm = dismod3.load_disease_model(disease)
for n in vars['sigma_alpha']:
try:
dm_na = dm.get_empirical_prior(full_name[data_type])['new_alpha']
model.parameters[data_type]['random_effects'][n.__name__] = dict(dist = dm_na[n.__name__]['dist'],
mu = dm_na[n.__name__]['mu'],
sigma = dm_na[n.__name__]['sigma'],
lower = dm_na[n.__name__]['lower'],
upper = dm_na[n.__name__]['upper'])
except:
model.parameters[data_type]['random_effects'][n.__name__] = dict(dist = 'TruncatedNormal',
mu = .05,
sigma = .03**-2,
lower = 0.01,
upper = 0.5)
# save fixed effects
emp_fe = pandas.read_csv('/home/j/Project/dismod/output/dm-%s/posterior/fe-%s-%s+%s+%s.csv'%(disease, data_type, geo_info(country,disease), sex, year), index_col=0)
for n, col in zip(vars['beta'], vars['X'].columns):
model.parameters[data_type]['fixed_effects'][col] = dict(dist = 'Constant',
mu = emp_fe.ix[col, 'mu_coeff'],
sigma = emp_fe.ix[col, 'sigma_coeff'])
def initialize_model():
### @export 'load model'
dm = dismod3.load_disease_model(19807)
### @export 'initialize model data'
dm.params['global_priors']['level_bounds']['excess_mortality'] = dict(lower=.1, upper=100.)
dm.params['global_priors']['increasing']['excess_mortality'] = dict(age_start=0, age_end=0)
dm.params['global_priors']['level_bounds']['relative_risk'] = dict(lower=0., upper=10000.)
for cv in dm.params['covariates']['Study_level']:
dm.params['covariates']['Study_level'][cv]['rate']['value'] = 0
for cv in dm.params['covariates']['Country_level']:
dm.params['covariates']['Country_level'][cv]['rate']['value'] = 0
level = .001
dm.params['sex_effect_prevalence'] = dict(mean=1, upper_ci=pl.exp(level *1.96), lower_ci=pl.exp(-level*1.96))
dm.params['time_effect_prevalence'] = dict(mean=1, upper_ci=pl.exp(level *1.96), lower_ci=pl.exp(-level*1.96))
dm.params['region_effect_prevalence'] = dict(std=level)
dm.clear_fit()
dm.clear_empirical_prior()
dismod3.neg_binom_model.covariate_hash = {}
return dm
def upload_fits(id):
""" Send results of cluster fits to dismod server
Parameters
----------
id : int
The model id number
Example
-------
>>> import fit_emp_prior
>>> fit_emp_prior.fit_emp_prior(2552, 'incidence')
>>> import upload_fits
>>> upload_fits.upload_fits(2552)
"""
# load disease model
dm = dismod3.load_disease_model(
id) # this merges together results from all fits
# save dta output
dir = dismod3.settings.JOB_WORKING_DIR % id # TODO: refactor into a function
#dm_to_dta(dm, '%s/regional_predictions' % dir)
# plot empirical priors (in a separate script, to run after all empirical priors are computed)
for effect in ['alpha', 'beta', 'gamma', 'delta']:
try:
dismod3.plotting.plot_empirical_prior_effects([dm], effect)
dm.savefig('dm-%d-emp-prior-%s.png' % (id, effect))
except Exception:
print 'failed to plot %s' % effect
# save table output
try:
dismod3.table.make_tables(dm)
except Exception, e:
print 'Failed to make table'
print e
def upload_fits(id):
""" Send results of cluster fits to dismod server
Parameters
----------
id : int
The model id number
Example
-------
>>> import fit_emp_prior
>>> fit_emp_prior.fit_emp_prior(2552, 'incidence')
>>> import upload_fits
>>> upload_fits.upload_fits(2552)
"""
# load disease model
dm = dismod3.load_disease_model(id) # this merges together results from all fits
# save dta output
dir = dismod3.settings.JOB_WORKING_DIR % id # TODO: refactor into a function
#dm_to_dta(dm, '%s/regional_predictions' % dir)
# plot empirical priors (in a separate script, to run after all empirical priors are computed)
for effect in ['alpha', 'beta', 'gamma', 'delta']:
try:
dismod3.plotting.plot_empirical_prior_effects([dm], effect)
dm.savefig('dm-%d-emp-prior-%s.png' % (id, effect))
except Exception:
print 'failed to plot %s' % effect
# save table output
try:
dismod3.table.make_tables(dm)
except Exception, e:
print 'Failed to make table'
print e
def fit_without_confrontation(id, region, sex, year):
""" Fit posterior of specified region/sex/year for specified model
without trying to integrate conflicting sources of data
Parameters
----------
id : int
The model id number for the job to fit
region : str
From dismod3.settings.gbd_regions, but clean()-ed
sex : str, from dismod3.settings.gbd_sexes
year : str, from dismod3.settings.gbd_years
"""
## load model
dm = dismod3.load_disease_model(id)
## separate out prevalence and relative-risk data
prev_data = [
d for d in dm.data
if dm.relevant_to(d, 'prevalence', region, year, sex)
]
rr_data = [
d for d in dm.data
if dm.relevant_to(d, 'relative-risk', region, year, sex)
]
dm.data = [d for d in dm.data if not d in prev_data and not d in rr_data]
### setup the generic disease model (without prevalence data)
import dismod3.gbd_disease_model as model
keys = dismod3.utils.gbd_keys(region_list=[region],
year_list=[year],
sex_list=[sex])
dm.calc_effective_sample_size(dm.data)
dm.vars = model.setup(dm, keys)
## override the birth prevalence prior, based on the withheld prevalence data
logit_C_0 = dm.vars[dismod3.utils.gbd_key_for('bins', region, year,
sex)]['initial']['logit_C_0']
assert len(prev_data) == 1, 'should be a single prevalance datum'
d = prev_data[0]
mu_logit_C_0 = mc.logit(dm.value_per_1(d) + dismod3.settings.NEARLY_ZERO)
lb, ub = dm.bounds_per_1(d)
sigma_logit_C_0 = (mc.logit(ub + dismod3.settings.NEARLY_ZERO) -
mc.logit(lb + dismod3.settings.NEARLY_ZERO)) / (2 *
1.96)
print 'mu_C_0_pri:', mc.invlogit(mu_logit_C_0)
print 'ui_C_0_pri:', lb, ub
# override the excess-mortality, based on the relative-risk data
mu_rr = 1.01 * np.ones(dismod3.settings.MAX_AGE)
sigma_rr = .01 * np.ones(dismod3.settings.MAX_AGE)
for d in rr_data:
mu_rr[d['age_start']:(d['age_end'] + 1)] = dm.value_per_1(d)
sigma_rr[d['age_start']:(d['age_end'] + 1)] = dm.se_per_1(d)
print 'mu_rr:', mu_rr.round(2)
#print 'sigma_rr:', sigma_rr.round(2)
log_f = dm.vars[dismod3.utils.gbd_key_for('excess-mortality', region, year,
sex)]['age_coeffs']
log_f_mesh = log_f.parents['gamma_mesh']
param_mesh = log_f.parents['param_mesh']
m_all = dm.vars[dismod3.utils.gbd_key_for('all-cause_mortality', region,
year, sex)]
mu_log_f = np.log((mu_rr - 1) * m_all)
sigma_log_f = 1 / ((mu_rr - 1) * m_all) * sigma_rr * m_all
print 'mu_log_f:', mu_log_f.round(2)[param_mesh]
print 'sigma_log_f:', sigma_log_f.round(2)[param_mesh]
### fit the model using Monte Carlo simulation (shoehorned into the MCMC framework of PyMC)
dm.mcmc = mc.MCMC(dm.vars)
dm.mcmc.use_step_method(SampleFromNormal,
logit_C_0,
mu=mu_logit_C_0,
tau=sigma_logit_C_0**-2)
dm.mcmc.use_step_method(SampleFromNormal,
log_f_mesh,
mu=mu_log_f[param_mesh],
tau=sigma_log_f[param_mesh]**-2)
for stoch in dm.mcmc.stochastics:
dm.mcmc.use_step_method(mc.NoStepper, stoch)
dm.mcmc.sample(1000, verbose=dismod3.settings.ON_SGE)
#print 'mu_C_0_post:', mc.invlogit(logit_C_0.stats()['mean']).round(2)
#print 'ui_C_0_post:', mc.invlogit(logit_C_0.stats()['95% HPD interval']).round(2)
#print 'mu_rr_post:', dm.vars[dismod3.utils.gbd_key_for('relative-risk', region, year, sex)]['rate_stoch'].stats()['mean'].round(2)
print 'mu_log_f_mesh_post:', log_f_mesh.stats()['mean'].round(2)
print 'mu_f_post:', dm.vars[dismod3.utils.gbd_key_for(
'excess-mortality', region, year,
sex)]['rate_stoch'].stats()['mean'].round(2)
for k in keys:
t, r, y, s = dismod3.utils.type_region_year_sex_from_key(k)
if t in [
'incidence', 'prevalence', 'remission', 'excess-mortality',
'mortality', 'prevalence_x_excess-mortality'
]:
dismod3.neg_binom_model.store_mcmc_fit(dm, k, dm.vars[k])
elif t in ['relative-risk', 'duration', 'incidence_x_duration']:
dismod3.normal_model.store_mcmc_fit(dm, k, dm.vars[k])
from fit_posterior import save_country_level_posterior
if str(year) == '2005': # also generate 2010 estimates
save_country_level_posterior(dm, region, 2010, sex,
['prevalence', 'remission'])
save_country_level_posterior(
dm, region, year, sex, ['prevalence', 'remission']
) #'prevalence incidence remission excess-mortality duration mortality relative-risk'.split())
# save results (do this last, because it removes things from the disease model that plotting function, etc, might need
keys = dismod3.utils.gbd_keys(region_list=[region],
year_list=[year],
sex_list=[sex])
dm.save('dm-%d-posterior-%s-%s-%s.json' % (dm.id, region, sex, year),
keys_to_save=keys)
return dm
def fit_emp_prior(
id,
param_type,
fast_fit=False,
generate_emp_priors=True,
zero_re=True,
alt_prior=False,
global_heterogeneity="Slightly",
):
""" Fit empirical prior of specified type for specified model
Parameters
----------
id : int
The model id number for the job to fit
param_type : str, one of incidence, prevalence, remission, excess-mortality, prevalence_x_excess-mortality
The disease parameter to generate empirical priors for
Example
-------
>>> import fit_emp_prior
>>> fit_emp_prior.fit_emp_prior(2552, 'incidence')
"""
dir = dismod3.settings.JOB_WORKING_DIR % id
## load the model from disk or from web
import simplejson as json
import data
reload(data)
dm = dismod3.load_disease_model(id)
try:
model = data.ModelData.load(dir)
print "loaded data from new format from %s" % dir
except (IOError, AssertionError):
model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
# model.save(dir)
print "loaded data from json, saved in new format for next time in %s" % dir
## next block fills in missing covariates with zero
for col in model.input_data.columns:
if col.startswith("x_"):
model.input_data[col] = model.input_data[col].fillna(0.0)
# also fill all covariates missing in output template with zeros
model.output_template = model.output_template.fillna(0)
# set all heterogeneity priors to Slightly for the global fit
for t in model.parameters:
if "heterogeneity" in model.parameters[t]:
model.parameters[t]["heterogeneity"] = global_heterogeneity
t = {
"incidence": "i",
"prevalence": "p",
"remission": "r",
"excess-mortality": "f",
"prevalence_x_excess-mortality": "pf",
}[param_type]
model.input_data = model.get_data(t)
if len(model.input_data) == 0:
print "No data for type %s, exiting" % param_type
return dm
### For testing:
## speed up computation by reducing number of knots
## model.parameters[t]['parameter_age_mesh'] = [0, 10, 20, 40, 60, 100]
## smooth Slightly, Moderately, or Very
## model.parameters[t]['smoothness'] = dict(age_start=0, age_end=100, amount='Very')
## speed up computation be reducing data size
## predict_area = 'super-region_0'
## predict_year=2005
## predict_sex='total'
## subtree = nx.traversal.bfs_tree(model.hierarchy, predict_area)
## relevant_rows = [i for i, r in model.input_data.T.iteritems() \
## if (r['area'] in subtree or r['area'] == 'all')\
## and (r['year_end'] >= 1997) \
## and r['sex'] in [predict_sex, 'total']]
## model.input_data = model.input_data.ix[relevant_rows]
# testing changes
# model.input_data['effective_sample_size'] = pl.minimum(1.e3, model.input_data['effective_sample_size'])
# missing_ess = pl.isnan(model.input_data['effective_sample_size'])
# model.input_data['effective_sample_size'][missing_ess] = 1.
# model.input_data['z_overdisperse'] = 1.
# print model.describe(t)
# model.input_data = model.input_data[model.input_data['area'].map(lambda x: x in nx.bfs_tree(model.hierarchy, 'super-region_5'))]
# model.input_data = model.input_data = model.input_data.drop(['x_LDI_id_Updated_7July2011'], axis=1)
# model.input_data = model.input_data.filter([model.input_data['x_nottroponinuse'] == 0.]
# model.input_data = model.input_data[:100]
## speed up output by not making predictions for empirical priors
# generate_emp_priors = False
print "fitting", t
model.vars += ism.age_specific_rate(
model,
t,
reference_area="all",
reference_sex="total",
reference_year="all",
mu_age=None,
mu_age_parent=None,
sigma_age_parent=None,
rate_type=(t == "rr") and "log_normal" or "neg_binom",
zero_re=zero_re,
)
# for backwards compatibility, should be removed eventually
dm.model = model
dm.vars = model.vars[t]
vars = dm.vars
if fast_fit:
dm.map, dm.mcmc = dismod3.fit.fit_asr(model, t, iter=101, burn=0, thin=1, tune_interval=100)
else:
dm.map, dm.mcmc = dismod3.fit.fit_asr(
model, t, iter=50000, burn=10000, thin=40, tune_interval=1000, verbose=True
)
stats = dm.vars["p_pred"].stats(batches=5)
dm.vars["data"]["mu_pred"] = stats["mean"]
dm.vars["data"]["sigma_pred"] = stats["standard deviation"]
stats = dm.vars["pi"].stats(batches=5)
dm.vars["data"]["mc_error"] = stats["mc error"]
dm.vars["data"]["residual"] = dm.vars["data"]["value"] - dm.vars["data"]["mu_pred"]
dm.vars["data"]["abs_residual"] = pl.absolute(dm.vars["data"]["residual"])
graphics.plot_fit(model, data_types=[t], ylab=["PY"], plot_config=(1, 1), fig_size=(8, 8))
if generate_emp_priors:
for a in [dismod3.utils.clean(a) for a in dismod3.settings.gbd_regions]:
print "generating empirical prior for %s" % a
for s in dismod3.settings.gbd_sexes:
for y in dismod3.settings.gbd_years:
key = dismod3.utils.gbd_key_for(param_type, a, y, s)
if t in model.parameters and "level_bounds" in model.parameters[t]:
lower = model.parameters[t]["level_bounds"]["lower"]
upper = model.parameters[t]["level_bounds"]["upper"]
else:
lower = 0
upper = pl.inf
emp_priors = covariate_model.predict_for(
model,
model.parameters[t],
"all",
"total",
"all",
a,
dismod3.utils.clean(s),
int(y),
alt_prior,
vars,
lower,
upper,
)
dm.set_mcmc("emp_prior_mean", key, emp_priors.mean(0))
if "eta" in vars:
N, A = emp_priors.shape # N samples, for A age groups
delta_trace = pl.transpose(
[pl.exp(vars["eta"].trace()) for _ in range(A)]
) # shape delta matrix to match prediction matrix
emp_prior_std = pl.sqrt(emp_priors.var(0) + (emp_priors ** 2 / delta_trace).mean(0))
else:
emp_prior_std = emp_priors.std(0)
dm.set_mcmc("emp_prior_std", key, emp_prior_std)
pl.plot(
model.parameters["ages"],
dm.get_mcmc("emp_prior_mean", key),
color="grey",
label=a,
zorder=-10,
alpha=0.5,
)
pl.savefig(dir + "/prior-%s.png" % param_type)
store_effect_coefficients(dm, vars, param_type)
# graphics.plot_one_ppc(vars, t)
# pl.savefig(dir + '/prior-%s-ppc.png'%param_type)
graphics.plot_acorr(model)
pl.savefig(dir + "/prior-%s-convergence.png" % param_type)
graphics.plot_trace(model)
pl.savefig(dir + "/prior-%s-trace.png" % param_type)
graphics.plot_one_effects(model, t)
pl.savefig(dir + "/prior-%s-effects.png" % param_type)
# save results (do this last, because it removes things from the disease model that plotting function, etc, might need
try:
dm.save("dm-%d-prior-%s.json" % (id, param_type))
except IOError, e:
print e
def validate_prior_similarity():
#dm = dismod3.load_disease_model(20945)
#dm.model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
#t = 'i'
#area, sex, year = 'europe_eastern', 'male', 2005
dm = dismod3.load_disease_model(20928)
dm.model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
t = 'p'
area, sex, year = 'sub-saharan_africa_central', 'male', 2005
# select data that is about areas in this region, recent years, and sex of male or total only
model = dm.model
subtree = nx.traversal.bfs_tree(model.hierarchy, area)
relevant_rows = [i for i, r in model.input_data.T.iteritems() \
if (r['area'] in subtree or r['area'] == 'all')\
and ((year == 2005 and r['year_end'] >= 1997) or r['year_start'] <= 1997) \
and r['sex'] in [sex, 'total']]
model.input_data = model.input_data.ix[relevant_rows]
# replace area 'all' with area
model.input_data['area'][model.input_data['area'] == 'all'] = area
for het in 'Slightly Moderately Very'.split():
dm.model.parameters[t]['parameter_age_mesh'] = [
0, 15, 20, 25, 35, 45, 55, 65, 75, 100
]
dm.model.parameters[t]['heterogeneity'] = het
setup_regional_model(dm, area, sex, year)
dm.vars = {}
dm.vars[t] = data_model.data_model(
t,
dm.model,
t,
root_area=area,
root_sex=sex,
root_year=year,
mu_age=None,
mu_age_parent=dm.emp_priors[t, 'mu'],
sigma_age_parent=dm.emp_priors[t, 'sigma'],
rate_type=(t == 'rr') and 'log_normal' or 'neg_binom')
fit_model.fit_data_model(dm.vars[t],
iter=10050,
burn=5000,
thin=50,
tune_interval=100)
#2graphics.plot_one_effects(dm.vars[t], t, dm.model.hierarchy)
#pl.title(het)
graphics.plot_convergence_diag(dm.vars[t])
pl.title(het)
#graphics.plot_one_ppc(dm.vars[t], t)
#pl.title(het)
graphics.plot_one_type(dm.model, dm.vars[t], dm.emp_priors, t)
pl.title(het)
pl.show()
return dm
pl.figure()
for iso in list(pl.unique(Y["Iso3"])):
pl.plot(Y[(Y["Iso3"] == iso) & (Y["Rate type"] == "prevalence")].filter(like="Draw").mean(1).__array__(), label=iso)
pl.semilogy([1], [1])
Z = Y[Y["Rate type"] == "prevalence"].groupby("Age").apply(weighted_age)
pl.plot(Z.mean(1).__array__(), color="red", linewidth=3, alpha=0.5, label="Inconsistent NA/ME")
pl.legend()
pl.axis([-5, 130, 1e-6, 2])
import dismod3
dm = dismod3.load_disease_model(19807)
import fit_posterior
fit_posterior.fit_posterior(dm, "north_africa_middle_east", "male", "2005", map_only=True)
X = pandas.read_csv(
"/var/tmp/dismod_working/test/dm-19807/posterior/dm-19807-north_africa_middle_east-male-2005.csv", index_col=None
)
pl.figure()
for iso in list(pl.unique(X["Iso3"])):
pl.plot(X[(X["Iso3"] == iso)].filter(like="Draw").mean(1).__array__(), label=iso)
pl.semilogy([1], [1])
Z = X.groupby("Age").apply(weighted_age)
plot(Z.mean(1).__array__(), color="red", linewidth=3, alpha=0.5, label="Inconsistent NA/ME")
for iso in list(pl.unique(Y['Iso3'])):
pl.plot(Y[(Y['Iso3']==iso)&(Y['Rate type']=='prevalence')].filter(like='Draw').mean(1).__array__(), label=iso)
pl.semilogy([1],[1])
Z = Y[Y['Rate type'] == 'prevalence'].groupby('Age').apply(weighted_age)
pl.plot(Z.mean(1).__array__(), color='red', linewidth=3, alpha=.5, label='Inconsistent NA/ME')
pl.legend()
pl.axis([-5,130,1e-6,2])
import dismod3
dm = dismod3.load_disease_model(19807)
import fit_posterior
fit_posterior.fit_posterior(dm, 'north_africa_middle_east', 'male', '2005', map_only=True)
X = pandas.read_csv('/var/tmp/dismod_working/test/dm-19807/posterior/dm-19807-north_africa_middle_east-male-2005.csv', index_col=None)
pl.figure()
for iso in list(pl.unique(X['Iso3'])):
pl.plot(X[(X['Iso3']==iso)].filter(like='Draw').mean(1).__array__(), label=iso)
pl.semilogy([1],[1])
Z = X.groupby('Age').apply(weighted_age)
plot(Z.mean(1).__array__(), color='red', linewidth=3, alpha=.5, label='Inconsistent NA/ME')
plot(dm.vars['prevalence+north_africa_middle_east+2005+male']['rate_stoch'].stats()['mean'], color='red', linewidth=3, alpha=.5, label='Mean of Consistent NA/ME')
def fit_all(id, consistent_empirical_prior=True, consistent_posterior=True,
posteriors_only=False, posterior_types='p i r', fast=False,
zero_re=True,
alt_prior=True,
global_heterogeneity='Slightly'):
""" Enqueues all jobs necessary to fit specified model
to the cluster
Parameters
----------
id : int
The model id number for the job to fit
Example
-------
>>> import fit_all
>>> fit_all.fit_all(2552)
"""
dir = dismod3.settings.JOB_WORKING_DIR % id # TODO: refactor into a function
try:
model = data.ModelData.load(dir)
print 'loaded data from new format from %s' % dir
# if we make it here, this model has already been run, so clean out the stdout/stderr dirs to make room for fresh messages
call_str = 'rm -rf %s/empirical_priors/stdout/* %s/empirical_priors/stderr/* %s/posterior/stdout/* %s/posterior/stderr/* %s/json/dm-*-*.json' % (dir, dir, dir, dir, dir)
print call_str
subprocess.call(call_str, shell=True)
# now load just the model, all previous fits are deleted
dm = dismod3.load_disease_model(id)
except (IOError, AssertionError):
print 'downloading disease model'
dm = dismod3.load_disease_model(id)
import simplejson as json
try:
model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
except Exception as e:
print e
print 'attempting to use old covariate format'
import old_cov_data
model = old_cov_data.from_gbd_jsons(json.loads(dm.to_json()))
model.save(dir)
print 'loaded data from json, saved in new format for next time in %s' % dir
def options(fast, zero_re, alt_prior, global_heterogeneity):
call_str = ''
call_str += ' --fast=%s'%fast
call_str += ' --zerore=%s'%zero_re
call_str += ' --altprior=%s'%alt_prior
call_str += ' --globalheterogeneity=%s'%global_heterogeneity
return call_str
o = '%s/empirical_priors/stdout/%d_running.txt' % (dir, id)
f = open(o, 'w')
import time
f.write('./run_on_cluster.sh fit_all.py --priorconsistent=%s --posteriorconsistent=%s %s\n' % (consistent_empirical_prior, consistent_posterior, options(fast, zero_re, alt_prior, global_heterogeneity)))
f.write('Enqueued model %d on cluster at %s' % (id, time.strftime('%c')))
f.close()
# fit empirical priors (by pooling data from all regions)
emp_names = []
if not posteriors_only:
if consistent_empirical_prior:
t = 'all'
o = '%s/empirical_priors/stdout/dismod_log_%s' % (dir, t)
e = '%s/empirical_priors/stderr/dismod_log_%s' % (dir, t)
name_str = '%s-%d' %(t[0], id)
emp_names.append(name_str)
if dismod3.settings.ON_SGE:
call_str = 'qsub -cwd -o %s -e %s ' % (o, e) \
+ '-N %s ' % name_str \
+ 'run_on_cluster.sh '
else:
call_str = 'python '
call_str += 'fit_world.py %d' % id
call_str += options(fast, zero_re, alt_prior, global_heterogeneity)
subprocess.call(call_str, shell=True)
else:
for t in ['excess-mortality', 'remission', 'incidence', 'prevalence']:
o = '%s/empirical_priors/stdout/dismod_log_%s' % (dir, t)
e = '%s/empirical_priors/stderr/dismod_log_%s' % (dir, t)
name_str = '%s-%d' %(t[0], id)
emp_names.append(name_str)
if dismod3.settings.ON_SGE:
call_str = 'qsub -cwd -o %s -e %s ' % (o, e) \
+ '-N %s ' % name_str \
+ 'run_on_cluster.sh '
else:
call_str = 'python '
call_str += 'fit_emp_prior.py %d -t %s' % (id, t)
call_str += options(fast, zero_re, alt_prior, global_heterogeneity)
subprocess.call(call_str, shell=True)
# directory to save the country level posterior csv files
temp_dir = dir + '/posterior/country_level_posterior_dm-' + str(id) + '/'
#fit each region/year/sex individually for this model
hold_str = '-hold_jid %s ' % ','.join(emp_names)
if posteriors_only:
hold_str = ''
post_names = []
for ii, r in enumerate(dismod3.gbd_regions):
for s in dismod3.gbd_sexes:
for y in dismod3.gbd_years:
k = '%s+%s+%s' % (dismod3.utils.clean(r), dismod3.utils.clean(s), y)
o = '%s/posterior/stdout/dismod_log_%s' % (dir, k)
e = '%s/posterior/stderr/dismod_log_%s' % (dir, k)
name_str = '%s%d%s%s%d' % (r[0], ii+1, s[0], str(y)[-1], id)
post_names.append(name_str)
if dismod3.settings.ON_SGE:
call_str = 'qsub -cwd -o %s -e %s ' % (o,e) \
+ hold_str \
+ '-N %s ' % name_str \
+ 'run_on_cluster.sh '
else:
call_str = 'python '
call_str += 'fit_posterior.py %d -r %s -s %s -y %s' % (id, dismod3.utils.clean(r), dismod3.utils.clean(s), y)
if not consistent_posterior:
call_str += ' --inconsistent=True --types="%s"' % posterior_types
if posteriors_only:
call_str += ' --onlyposterior=True'
if fast:
call_str += ' --fast=true'
call_str += ' --zerore=%s'%zero_re
subprocess.call(call_str, shell=True)
# after all posteriors have finished running, upload disease model json
hold_str = '-hold_jid %s ' % ','.join(post_names)
o = '%s/empirical_priors/stdout/%d_upload.txt' % (dir, id)
e = '%s/empirical_priors/stderr/%d_upload.txt' % (dir, id)
if dismod3.settings.ON_SGE:
call_str = 'qsub -cwd -o %s -e %s ' % (o,e) \
+ hold_str \
+ '-N upld-%s ' % id \
+ 'run_on_cluster.sh '
else:
call_str = 'python '
call_str += 'upload_fits.py %d' % id
subprocess.call(call_str, shell=True)
return dm
import dismod3
import book_graphics
reload(book_graphics)
results = {}
n_pred = 10000
iter = 20000
burn = 10000
thin = 10
# set font
book_graphics.set_font()
### @export 'data'
# TODO: migrate data into a csv, load with pandas
dm = dismod3.load_disease_model(15630)
dm.calc_effective_sample_size(dm.data)
some_data = ([d for d in dm.data
if d['data_type'] == 'prevalence data'
and d['sex'] == 'male'
and 15 <= d['age_start'] < 20
and d['age_end'] == 99
and d['effective_sample_size'] > 1])
countries = pl.unique([s['region'] for s in some_data])
min_year = min([s['year_start'] for s in some_data])
max_year = max([s['year_end'] for s in some_data])
cy = ['%s-%d'%(s['region'], s['year_start']) for s in some_data]
n = pl.array([s['effective_sample_size'] for s in some_data])
r = pl.array([dm.value_per_1(s) for s in some_data])
import pylab as pl
import pymc as mc
import dismod3
import book_graphics
reload(book_graphics)
# set font
book_graphics.set_font()
results = {}
### @export 'data'
#dm = dismod3.load_disease_model(15596) # epilipsy
dm = dismod3.load_disease_model(16240) # af
data = dm.filter_data('prevalence+all+all+all')
hist = pl.zeros((101,101))
for d in data:
hist[d['age_start'], d['age_end']] += 1
most_freq_cnt = hist.max()
rows_total = len(data)
### @export 'scatter-prevalence-age-groups'
pl.figure(**book_graphics.half_page_params)
#pl.subplot(1,2,2)
for a_0 in range(101):
def measure_fit(id, condition):
"""
Determine the RMSE, MAE, and Coverage of the fit stored in model specified by id
"""
print 'downloading model %d' % id
sys.stdout.flush()
dm = dismod3.load_disease_model(id)
#print 'loading gold-standard data'
gold_data = [
d
for d in csv.DictReader(open(OUTPUT_PATH + '%s_gold.tsv' % condition),
dialect='excel-tab')
]
#print 'comparing values'
abs_err = dict(incidence=[],
prevalence=[],
remission=[],
duration=[],
incidence_x_duration=[])
rel_err = dict(incidence=[],
prevalence=[],
remission=[],
duration=[],
incidence_x_duration=[])
coverage = dict(incidence=[],
prevalence=[],
remission=[],
duration=[],
incidence_x_duration=[])
for metric in [abs_err, rel_err, coverage]:
metric['excess mortality'] = []
for d in gold_data:
est = predict('mean', dm, d)
lb = predict('lower_ui', dm, d)
ub = predict('upper_ui', dm, d)
if est < 0:
continue
val = float(d['Truth'])
err = val - est
#if d['Age Start'] <= 50:
# continue
t = d['Parameter'].replace(' data', '')
abs_err[t].append(err)
if val > 0.:
rel_err[t].append(100 * err / val)
coverage[t].append(val >= lb and val <= ub)
for k in abs_err:
print '%s abs RMSE = %f' % (k, np.sqrt(np.mean(
np.array(abs_err[k])**2)))
print '%s abs MAE = %f' % (k, np.median(np.abs(abs_err[k])))
print
for k in rel_err:
print '%s rel pct RMSE = %f' % (
k, np.sqrt(np.mean(np.array(rel_err[k])**2)))
print '%s rel pct MAE = %f' % (k, np.median(np.abs(rel_err[k])))
print
for k in coverage:
print '%s coverage = %f' % (k, np.sum(coverage[k]) * 100. /
len(coverage[k]))
print
k = 'incidence_x_duration'
print '%s rel pct MAE =\t%f' % (k, np.median(np.abs(rel_err[k])))
return np.median(np.abs(rel_err[k]))
# add estimate value as a column in the gold data tsv, for looking
# in more detail with a spreadsheet or different code
col_names = sorted(set(gold_data[0].keys()) | set(['Estimate Value']))
f_file = open(OUTPUT_PATH + '%s_gold.tsv' % condition, 'w')
csv_f = csv.writer(f_file, dialect='excel-tab')
csv_f.writerow(col_names)
csv_f = csv.DictWriter(f_file, col_names, dialect='excel-tab')
for d in gold_data:
csv_f.writerow(d)
f_file.close()
def fit_posterior(id, region, sex, year):
""" Fit posterior of specified region/sex/year for specified model
Parameters
----------
id : int
The model id number for the job to fit
region : str
From dismod3.settings.gbd_regions, but clean()-ed
sex : str, from dismod3.settings.gbd_sexes
year : str, from dismod3.settings.gbd_years
Example
-------
>>> import fit_posterior
>>> fit_posterior.fit_posterior(2552, 'asia_east', 'male', '2005')
"""
#print 'updating job status on server'
#dismod3.log_job_status(id, 'posterior', '%s--%s--%s' % (region, sex, year), 'Running')
dm = dismod3.load_disease_model(id)
#dm.data = [] # for testing, remove all data
keys = dismod3.utils.gbd_keys(region_list=[region], year_list=[year], sex_list=[sex])
# fit the model
dir = dismod3.settings.JOB_WORKING_DIR % id
import dismod3.gbd_disease_model as model
model.fit(dm, method='map', keys=keys, verbose=1) ## first generate decent initial conditions
## then sample the posterior via MCMC
model.fit(dm, method='mcmc', keys=keys, iter=50000, thin=25, burn=25000, verbose=1,
dbname='%s/posterior/pickle/dm-%d-posterior-%s-%s-%s.pickle' % (dir, id, region, sex, year))
# generate plots of results
dismod3.tile_plot_disease_model(dm, keys, defaults={})
dm.savefig('dm-%d-posterior-%s.png' % (id, '+'.join(['all', region, sex, year]))) # TODO: refactor naming into its own function (disease_json.save_image perhaps)
for param_type in dismod3.settings.output_data_types:
keys = dismod3.utils.gbd_keys(region_list=[region], year_list=[year], sex_list=[sex], type_list=[param_type])
dismod3.tile_plot_disease_model(dm, keys, defaults={})
dm.savefig('dm-%d-posterior-%s-%s-%s-%s.png' % (id, dismod3.utils.clean(param_type), region, sex, year)) # TODO: refactor naming into its own function
# summarize fit quality graphically, as well as parameter posteriors
for k in dismod3.utils.gbd_keys(region_list=[region], year_list=[year], sex_list=[sex]):
if dm.vars[k].get('data'):
dismod3.plotting.plot_posterior_predicted_checks(dm, k)
dm.savefig('dm-%d-check-%s.png' % (dm.id, k))
# save results (do this last, because it removes things from the disease model that plotting function, etc, might need
keys = dismod3.utils.gbd_keys(region_list=[region], year_list=[year], sex_list=[sex])
dm.save('dm-%d-posterior-%s-%s-%s.json' % (id, region, sex, year), keys_to_save=keys)
# make a rate_type_list
rate_type_list = ['incidence', 'prevalence', 'remission', 'excess-mortality',
'mortality', 'relative-risk', 'duration', 'incidence_x_duration']
# save country level posterior
save_country_level_posterior(dm, region, year, sex)
# update job status file
#print 'updating job status on server'
#dismod3.log_job_status(id, 'posterior',
# '%s--%s--%s' % (region, sex, year), 'Completed')
return dm
def main():
import optparse
usage = 'usage: %prog [options] disease_model_id'
parser = optparse.OptionParser(usage)
parser.add_option(
'-s',
'--sex',
default='male',
help=
'only estimate given sex (valid settings ``male``, ``female``, ``all``)'
)
parser.add_option(
'-y',
'--year',
default='2005',
help=
'only estimate given year (valid settings ``1990``, ``2005``, ``2010``)'
)
parser.add_option('-r',
'--region',
default='australasia',
help='only estimate given GBD Region')
parser.add_option('-f', '--fast', default='False', help='use MAP only')
parser.add_option('-i',
'--inconsistent',
default='False',
help='use inconsistent model for posteriors')
parser.add_option(
'-t',
'--types',
default='p i r',
help='with rate types to fit (only used if inconsistent=true)')
parser.add_option('-z',
'--zerore',
default='true',
help='enforce zero constraint on random effects')
parser.add_option('-o',
'--onlyposterior',
default='False',
help='skip empirical prior phase')
(options, args) = parser.parse_args()
if len(args) != 1:
parser.error('incorrect number of arguments')
try:
id = int(args[0])
except ValueError:
parser.error('disease_model_id must be an integer')
dm = dismod3.load_disease_model(id)
# set model id to passed-in id (should not be necessary)
dm.id = id
assert id == dm.id, 'model id should equal parameter id'
dm = fit_posterior(
dm,
options.region,
options.sex,
options.year,
fast_fit=options.fast.lower() == 'true',
inconsistent_fit=options.inconsistent.lower() == 'true',
params_to_fit=options.types.split(),
posteriors_only=(options.onlyposterior.lower() == 'true'),
zero_re=options.zerore.lower() == 'true')
return dm
def fit_emp_prior(id,
param_type,
fast_fit=False,
generate_emp_priors=True,
zero_re=True,
alt_prior=False,
global_heterogeneity='Slightly'):
""" Fit empirical prior of specified type for specified model
Parameters
----------
id : int
The model id number for the job to fit
param_type : str, one of incidence, prevalence, remission, excess-mortality, prevalence_x_excess-mortality
The disease parameter to generate empirical priors for
Example
-------
>>> import fit_emp_prior
>>> fit_emp_prior.fit_emp_prior(2552, 'incidence')
"""
dir = dismod3.settings.JOB_WORKING_DIR % id
## load the model from disk or from web
import simplejson as json
import data
reload(data)
dm = dismod3.load_disease_model(id)
try:
model = data.ModelData.load(dir)
print 'loaded data from new format from %s' % dir
except (IOError, AssertionError):
model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
#model.save(dir)
print 'loaded data from json, saved in new format for next time in %s' % dir
## next block fills in missing covariates with zero
for col in model.input_data.columns:
if col.startswith('x_'):
model.input_data[col] = model.input_data[col].fillna(0.)
# also fill all covariates missing in output template with zeros
model.output_template = model.output_template.fillna(0)
# set all heterogeneity priors to Slightly for the global fit
for t in model.parameters:
if 'heterogeneity' in model.parameters[t]:
model.parameters[t]['heterogeneity'] = global_heterogeneity
t = {
'incidence': 'i',
'prevalence': 'p',
'remission': 'r',
'excess-mortality': 'f',
'prevalence_x_excess-mortality': 'pf'
}[param_type]
model.input_data = model.get_data(t)
if len(model.input_data) == 0:
print 'No data for type %s, exiting' % param_type
return dm
### For testing:
## speed up computation by reducing number of knots
## model.parameters[t]['parameter_age_mesh'] = [0, 10, 20, 40, 60, 100]
## smooth Slightly, Moderately, or Very
## model.parameters[t]['smoothness'] = dict(age_start=0, age_end=100, amount='Very')
## speed up computation be reducing data size
## predict_area = 'super-region_0'
## predict_year=2005
## predict_sex='total'
## subtree = nx.traversal.bfs_tree(model.hierarchy, predict_area)
## relevant_rows = [i for i, r in model.input_data.T.iteritems() \
## if (r['area'] in subtree or r['area'] == 'all')\
## and (r['year_end'] >= 1997) \
## and r['sex'] in [predict_sex, 'total']]
## model.input_data = model.input_data.ix[relevant_rows]
# testing changes
#model.input_data['effective_sample_size'] = pl.minimum(1.e3, model.input_data['effective_sample_size'])
#missing_ess = pl.isnan(model.input_data['effective_sample_size'])
#model.input_data['effective_sample_size'][missing_ess] = 1.
#model.input_data['z_overdisperse'] = 1.
#print model.describe(t)
#model.input_data = model.input_data[model.input_data['area'].map(lambda x: x in nx.bfs_tree(model.hierarchy, 'super-region_5'))]
#model.input_data = model.input_data = model.input_data.drop(['x_LDI_id_Updated_7July2011'], axis=1)
#model.input_data = model.input_data.filter([model.input_data['x_nottroponinuse'] == 0.]
#model.input_data = model.input_data[:100]
## speed up output by not making predictions for empirical priors
#generate_emp_priors = False
print 'fitting', t
model.vars += ism.age_specific_rate(model,
t,
reference_area='all',
reference_sex='total',
reference_year='all',
mu_age=None,
mu_age_parent=None,
sigma_age_parent=None,
rate_type=(t == 'rr') and 'log_normal'
or 'neg_binom',
zero_re=zero_re)
# for backwards compatibility, should be removed eventually
dm.model = model
dm.vars = model.vars[t]
vars = dm.vars
if fast_fit:
dm.map, dm.mcmc = dismod3.fit.fit_asr(model,
t,
iter=101,
burn=0,
thin=1,
tune_interval=100)
else:
dm.map, dm.mcmc = dismod3.fit.fit_asr(model,
t,
iter=50000,
burn=10000,
thin=40,
tune_interval=1000,
verbose=True)
stats = dm.vars['p_pred'].stats(batches=5)
dm.vars['data']['mu_pred'] = stats['mean']
dm.vars['data']['sigma_pred'] = stats['standard deviation']
stats = dm.vars['pi'].stats(batches=5)
dm.vars['data']['mc_error'] = stats['mc error']
dm.vars['data'][
'residual'] = dm.vars['data']['value'] - dm.vars['data']['mu_pred']
dm.vars['data']['abs_residual'] = pl.absolute(dm.vars['data']['residual'])
graphics.plot_fit(model,
data_types=[t],
ylab=['PY'],
plot_config=(1, 1),
fig_size=(8, 8))
if generate_emp_priors:
for a in [
dismod3.utils.clean(a) for a in dismod3.settings.gbd_regions
]:
print 'generating empirical prior for %s' % a
for s in dismod3.settings.gbd_sexes:
for y in dismod3.settings.gbd_years:
key = dismod3.utils.gbd_key_for(param_type, a, y, s)
if t in model.parameters and 'level_bounds' in model.parameters[
t]:
lower = model.parameters[t]['level_bounds']['lower']
upper = model.parameters[t]['level_bounds']['upper']
else:
lower = 0
upper = pl.inf
emp_priors = covariate_model.predict_for(
model, model.parameters[t], 'all', 'total', 'all', a,
dismod3.utils.clean(s), int(y), alt_prior, vars, lower,
upper)
dm.set_mcmc('emp_prior_mean', key, emp_priors.mean(0))
if 'eta' in vars:
N, A = emp_priors.shape # N samples, for A age groups
delta_trace = pl.transpose([
pl.exp(vars['eta'].trace()) for _ in range(A)
]) # shape delta matrix to match prediction matrix
emp_prior_std = pl.sqrt(
emp_priors.var(0) +
(emp_priors**2 / delta_trace).mean(0))
else:
emp_prior_std = emp_priors.std(0)
dm.set_mcmc('emp_prior_std', key, emp_prior_std)
pl.plot(model.parameters['ages'],
dm.get_mcmc('emp_prior_mean', key),
color='grey',
label=a,
zorder=-10,
alpha=.5)
pl.savefig(dir + '/prior-%s.png' % param_type)
store_effect_coefficients(dm, vars, param_type)
#graphics.plot_one_ppc(vars, t)
#pl.savefig(dir + '/prior-%s-ppc.png'%param_type)
graphics.plot_acorr(model)
pl.savefig(dir + '/prior-%s-convergence.png' % param_type)
graphics.plot_trace(model)
pl.savefig(dir + '/prior-%s-trace.png' % param_type)
graphics.plot_one_effects(model, t)
pl.savefig(dir + '/prior-%s-effects.png' % param_type)
# save results (do this last, because it removes things from the disease model that plotting function, etc, might need
try:
dm.save('dm-%d-prior-%s.json' % (id, param_type))
except IOError, e:
print e
def measure_fit(id, condition):
"""
Determine the RMSE, MAE, and Coverage of the fit stored in model specified by id
"""
print 'downloading model %d' % id
sys.stdout.flush()
dm = dismod3.load_disease_model(id)
#print 'loading gold-standard data'
gold_data = [d for d in csv.DictReader(open(OUTPUT_PATH + '%s_gold.tsv' % condition), dialect='excel-tab')]
#print 'comparing values'
abs_err = dict(incidence=[], prevalence=[], remission=[], duration=[], incidence_x_duration=[])
rel_err = dict(incidence=[], prevalence=[], remission=[], duration=[], incidence_x_duration=[])
coverage = dict(incidence=[], prevalence=[], remission=[], duration=[], incidence_x_duration=[])
for metric in [abs_err, rel_err, coverage]:
metric['excess mortality'] = []
for d in gold_data:
est = predict('mean', dm, d)
lb = predict('lower_ui', dm, d)
ub = predict('upper_ui', dm, d)
if est < 0:
continue
val = float(d['Truth'])
err = val - est
#if d['Age Start'] <= 50:
# continue
t = d['Parameter'].replace(' data', '')
abs_err[t].append(err)
if val > 0.:
rel_err[t].append(100 * err / val)
coverage[t].append(val >= lb and val <= ub)
for k in abs_err:
print '%s abs RMSE = %f' % (k, np.sqrt(np.mean(np.array(abs_err[k])**2)))
print '%s abs MAE = %f' % (k, np.median(np.abs(abs_err[k])))
print
for k in rel_err:
print '%s rel pct RMSE = %f' % (k, np.sqrt(np.mean(np.array(rel_err[k])**2)))
print '%s rel pct MAE = %f' % (k, np.median(np.abs(rel_err[k])))
print
for k in coverage:
print '%s coverage = %f' % (k, np.sum(coverage[k]) * 100. / len(coverage[k]))
print
k = 'incidence_x_duration'
print '%s rel pct MAE =\t%f' % (k, np.median(np.abs(rel_err[k])))
return np.median(np.abs(rel_err[k]))
# add estimate value as a column in the gold data tsv, for looking
# in more detail with a spreadsheet or different code
col_names = sorted(set(gold_data[0].keys()) | set(['Estimate Value']))
f_file = open(OUTPUT_PATH + '%s_gold.tsv' % condition, 'w')
csv_f = csv.writer(f_file, dialect='excel-tab')
csv_f.writerow(col_names)
csv_f = csv.DictWriter(f_file, col_names, dialect='excel-tab')
for d in gold_data:
csv_f.writerow(d)
f_file.close()
print '\nparam prediction bias: %.5f, MARE: %.3f, coverage: %.2f' % (model.mu['abs_err'].mean(),
pl.median(pl.absolute(model.mu['rel_err'].dropna())),
model.mu['covered?'].mean())
print
data_simulation.initialize_results(model)
data_simulation.add_to_results(model, 'mu')
data_simulation.finalize_results(model)
print model.results
return model
if __name__ == '__main__':
region, sex, year = 'north_america_high_income', 'male', 1990
import fit_posterior, upload_fits
import data
import simplejson as json
## load the model from disk or from web
dm = dismod3.load_disease_model(24842)
dm.model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
data = upload_fits.merge_data_csvs(24842)
dm.model.input_data['mu_pred'] = data['mu_pred']
simulate_data(dm, region, sex, year)
fit_simulated(dm, region, sex, year)
store_results(dm, region, sex, year)
import pylab as pl
import pymc as mc
import dismod3
import book_graphics
reload(book_graphics)
# set font
book_graphics.set_font()
results = {}
### @export 'data'
#dm = dismod3.load_disease_model(15596) # epilipsy
dm = dismod3.load_disease_model(16240) # af
data = dm.filter_data('prevalence+all+all+all')
hist = pl.zeros((101, 101))
for d in data:
hist[d['age_start'], d['age_end']] += 1
most_freq_cnt = hist.max()
rows_total = len(data)
### @export 'scatter-prevalence-age-groups'
pl.figure(**book_graphics.half_page_params)
#pl.subplot(1,2,2)
for a_0 in range(101):
def fit_world(id, fast_fit=False, zero_re=True, alt_prior=False, global_heterogeneity='Slightly'):
""" Fit consistent for all data in world
Parameters
----------
id : int
The model id number for the job to fit
Example
-------
>>> import fit_world
>>> dm = fit_world.dismod3.load_disease_model(1234)
>>> fit_world.fit_world(dm)
"""
dir = dismod3.settings.JOB_WORKING_DIR % id
## load the model from disk or from web
import simplejson as json
import data
reload(data)
try:
model = data.ModelData.load(dir)
print 'loaded data from new format from %s' % dir
dm = dismod3.load_disease_model(id)
except (IOError, AssertionError):
dm = dismod3.load_disease_model(id)
model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
try:
model.save(dir)
print 'loaded data from json, saved in new format for next time in %s' % dir
except IOError:
print 'loaded data from json, failed to save in new format'
## next block fills in missing covariates with zero
for col in model.input_data.columns:
if col.startswith('x_'):
model.input_data[col] = model.input_data[col].fillna(0.)
# also fill all covariates missing in output template with zeros
model.output_template = model.output_template.fillna(0)
# set all heterogeneity priors to Slightly for the global fit
for t in model.parameters:
if 'heterogeneity' in model.parameters[t]:
model.parameters[t]['heterogeneity'] = global_heterogeneity
### For testing:
## speed up computation by reducing number of knots
## for t in 'irf':
## model.parameters[t]['parameter_age_mesh'] = [0, 100]
model.vars += dismod3.ism.consistent(model,
reference_area='all',
reference_sex='total',
reference_year='all',
priors={},
zero_re=zero_re)
## fit model to data
if fast_fit:
dm.map, dm.mcmc = dismod3.fit.fit_consistent(model, 105, 0, 1, 100)
else:
dm.map, dm.mcmc = dismod3.fit.fit_consistent(model, iter=50000, burn=10000, thin=40, tune_interval=1000, verbose=True)
dm.model = model
# borrow strength to inform sigma_alpha between rate types post-hoc
types_with_re = ['rr', 'f', 'i', 'm', 'smr', 'p', 'r', 'pf', 'm_with', 'X']
## first calculate sigma_alpha_bar from posterior draws from each alpha
alpha_vals = []
for type in types_with_re:
if 'alpha' in model.vars[type]:
for alpha_i in model.vars[type]['alpha']:
alpha_vals += [a for a in alpha_i.trace() if a != 0] # remove zeros because areas with no siblings are included for convenience but are pinned to zero
## then blend sigma_alpha_i and sigma_alpha_bar for each sigma_alpha_i
if len(alpha_vals) > 0:
sigma_alpha_bar = pl.std(alpha_vals)
for type in types_with_re:
if 'sigma_alpha' in model.vars[type]:
for sigma_alpha_i in model.vars[type]['sigma_alpha']:
cur_val = sigma_alpha_i.trace()
sigma_alpha_i.trace._trace[0] = (cur_val + sigma_alpha_bar) * pl.ones_like(sigma_alpha_i.trace._trace[0])
for t in 'p i r f rr pf m_with'.split():
param_type = dict(i='incidence', r='remission', f='excess-mortality', p='prevalence', rr='relative-risk', pf='prevalence_x_excess-mortality', m_with='mortality')[t]
#graphics.plot_one_type(model, model.vars[t], {}, t)
for a in [dismod3.utils.clean(a) for a in dismod3.settings.gbd_regions]:
print 'generating empirical prior for %s' % a
for s in dismod3.settings.gbd_sexes:
for y in dismod3.settings.gbd_years:
key = dismod3.utils.gbd_key_for(param_type, a, y, s)
if t in model.parameters and 'level_bounds' in model.parameters[t]:
lower=model.parameters[t]['level_bounds']['lower']
upper=model.parameters[t]['level_bounds']['upper']
else:
lower=0
upper=pl.inf
emp_priors = covariate_model.predict_for(model,
model.parameters.get(t, {}),
'all', 'total', 'all',
a, dismod3.utils.clean(s), int(y),
alt_prior,
model.vars[t], lower, upper)
dm.set_mcmc('emp_prior_mean', key, emp_priors.mean(0))
if 'eta' in model.vars[t]:
N,A = emp_priors.shape # N samples, for A age groups
delta_trace = pl.transpose([pl.exp(model.vars[t]['eta'].trace()) for _ in range(A)]) # shape delta matrix to match prediction matrix
emp_prior_std = pl.sqrt(emp_priors.var(0) + (emp_priors**2 / delta_trace).mean(0))
else:
emp_prior_std = emp_priors.std(0)
dm.set_mcmc('emp_prior_std', key, emp_prior_std)
from fit_emp_prior import store_effect_coefficients
store_effect_coefficients(dm, model.vars[t], param_type)
if 'p_pred' in model.vars[t]:
graphics.plot_one_ppc(model, t)
pl.savefig(dir + '/prior-%s-ppc.png'%param_type)
if 'p_pred' in model.vars[t] or 'lb' in model.vars[t]:
graphics.plot_one_effects(model, t)
pl.savefig(dir + '/prior-%s-effects.png'%param_type)
for t in 'i r f p rr pf X m_with smr'.split():
fname = dir + '/empirical_priors/data-%s.csv'%t
print 'saving tables for', t, 'to', fname
if 'data' in model.vars[t] and 'p_pred' in model.vars[t]:
stats = model.vars[t]['p_pred'].stats(batches=5)
model.vars[t]['data']['mu_pred'] = stats['mean']
model.vars[t]['data']['sigma_pred'] = stats['standard deviation']
stats = model.vars[t]['pi'].stats(batches=5)
model.vars[t]['data']['mc_error'] = stats['mc error']
model.vars[t]['data']['residual'] = model.vars[t]['data']['value'] - model.vars[t]['data']['mu_pred']
model.vars[t]['data']['abs_residual'] = pl.absolute(model.vars[t]['data']['residual'])
#if 'delta' in model.vars[t]:
# model.vars[t]['data']['logp'] = [mc.negative_binomial_like(n*p_obs, n*p_pred, n*p_pred*d) for n, p_obs, p_pred, d \
# in zip(model.vars[t]['data']['effective_sample_size'],
# model.vars[t]['data']['value'],
# model.vars[t]['data']['mu_pred'],
# pl.atleast_1d(model.vars[t]['delta'].stats()['mean']))]
model.vars[t]['data'].to_csv(fname)
graphics.plot_fit(model)
pl.savefig(dir + '/prior.png')
graphics.plot_acorr(model)
pl.savefig(dir + '/prior-convergence.png')
graphics.plot_trace(model)
pl.savefig(dir + '/prior-trace.png')
# save results (do this last, because it removes things from the disease model that plotting function, etc, might need
try:
dm.save('dm-%d-prior-%s.json' % (dm.id, 'all'))
except IOError, e:
print e
import book_graphics
reload(book_graphics)
results = {}
### @export 'data'
region = 'north_america_high_income'
sex = 'female'
year = '2005'
heterogeneity = ['Slightly', 'Very']
for ii in range(2):
# load model
dm = dismod3.load_disease_model(16370)
# set expert priors and other model parameters
dm.set_param_age_mesh([0, 15, 20, 25, 30, 35, 40, 45, 50, 100])
dm.params['global_priors']['level_value']['incidence']['age_before'] = 15
dm.params['global_priors']['level_value']['incidence']['age_after'] = 50
dm.params['global_priors']['smoothness']['incidence']['age_start'] = 15
dm.params['global_priors']['level_value']['remission']['age_before'] = 40
dm.params['global_priors']['level_bounds']['remission']['upper'] = 10.
dm.params['global_priors']['level_value']['excess_mortality'][
'age_before'] = 101
dm.params['global_priors']['level_value']['prevalence']['age_before'] = 15
def fit_without_confrontation(id, region, sex, year):
""" Fit posterior of specified region/sex/year for specified model
without trying to integrate conflicting sources of data
Parameters
----------
id : int
The model id number for the job to fit
region : str
From dismod3.settings.gbd_regions, but clean()-ed
sex : str, from dismod3.settings.gbd_sexes
year : str, from dismod3.settings.gbd_years
"""
## load model
dm = dismod3.load_disease_model(id)
## separate out prevalence and relative-risk data
prev_data = [d for d in dm.data if dm.relevant_to(d, 'prevalence', region, year, sex)]
rr_data = [d for d in dm.data if dm.relevant_to(d, 'relative-risk', region, year, sex)]
dm.data = [d for d in dm.data if not d in prev_data and not d in rr_data]
### setup the generic disease model (without prevalence data)
import dismod3.gbd_disease_model as model
keys = dismod3.utils.gbd_keys(region_list=[region], year_list=[year], sex_list=[sex])
dm.calc_effective_sample_size(dm.data)
dm.vars = model.setup(dm, keys)
## override the birth prevalence prior, based on the withheld prevalence data
logit_C_0 = dm.vars[dismod3.utils.gbd_key_for('bins', region, year, sex)]['initial']['logit_C_0']
assert len(prev_data) == 1, 'should be a single prevalance datum'
d = prev_data[0]
mu_logit_C_0 = mc.logit(dm.value_per_1(d)+dismod3.settings.NEARLY_ZERO)
lb, ub = dm.bounds_per_1(d)
sigma_logit_C_0 = (mc.logit(ub+dismod3.settings.NEARLY_ZERO) - mc.logit(lb+dismod3.settings.NEARLY_ZERO)) / (2 * 1.96)
print 'mu_C_0_pri:', mc.invlogit(mu_logit_C_0)
print 'ui_C_0_pri:', lb, ub
# override the excess-mortality, based on the relative-risk data
mu_rr = 1.01*np.ones(dismod3.settings.MAX_AGE)
sigma_rr = .01*np.ones(dismod3.settings.MAX_AGE)
for d in rr_data:
mu_rr[d['age_start']:(d['age_end']+1)] = dm.value_per_1(d)
sigma_rr[d['age_start']:(d['age_end']+1)] = dm.se_per_1(d)
print 'mu_rr:', mu_rr.round(2)
#print 'sigma_rr:', sigma_rr.round(2)
log_f = dm.vars[dismod3.utils.gbd_key_for('excess-mortality', region, year, sex)]['age_coeffs']
log_f_mesh = log_f.parents['gamma_mesh']
param_mesh = log_f.parents['param_mesh']
m_all = dm.vars[dismod3.utils.gbd_key_for('all-cause_mortality', region, year, sex)]
mu_log_f = np.log((mu_rr-1) * m_all)
sigma_log_f = 1 / ((mu_rr-1) * m_all) * sigma_rr * m_all
print 'mu_log_f:', mu_log_f.round(2)[param_mesh]
print 'sigma_log_f:', sigma_log_f.round(2)[param_mesh]
### fit the model using Monte Carlo simulation (shoehorned into the MCMC framework of PyMC)
dm.mcmc = mc.MCMC(dm.vars)
dm.mcmc.use_step_method(SampleFromNormal, logit_C_0, mu=mu_logit_C_0, tau=sigma_logit_C_0**-2)
dm.mcmc.use_step_method(SampleFromNormal, log_f_mesh, mu=mu_log_f[param_mesh], tau=sigma_log_f[param_mesh]**-2)
for stoch in dm.mcmc.stochastics:
dm.mcmc.use_step_method(mc.NoStepper, stoch)
dm.mcmc.sample(1000, verbose=dismod3.settings.ON_SGE)
#print 'mu_C_0_post:', mc.invlogit(logit_C_0.stats()['mean']).round(2)
#print 'ui_C_0_post:', mc.invlogit(logit_C_0.stats()['95% HPD interval']).round(2)
#print 'mu_rr_post:', dm.vars[dismod3.utils.gbd_key_for('relative-risk', region, year, sex)]['rate_stoch'].stats()['mean'].round(2)
print 'mu_log_f_mesh_post:', log_f_mesh.stats()['mean'].round(2)
print 'mu_f_post:', dm.vars[dismod3.utils.gbd_key_for('excess-mortality', region, year, sex)]['rate_stoch'].stats()['mean'].round(2)
for k in keys:
t,r,y,s = dismod3.utils.type_region_year_sex_from_key(k)
if t in ['incidence', 'prevalence', 'remission', 'excess-mortality', 'mortality', 'prevalence_x_excess-mortality']:
dismod3.neg_binom_model.store_mcmc_fit(dm, k, dm.vars[k])
elif t in ['relative-risk', 'duration', 'incidence_x_duration']:
dismod3.normal_model.store_mcmc_fit(dm, k, dm.vars[k])
from fit_posterior import save_country_level_posterior
if str(year) == '2005': # also generate 2010 estimates
save_country_level_posterior(dm, region, 2010, sex, ['prevalence', 'remission'])
save_country_level_posterior(dm, region, year, sex, ['prevalence', 'remission']) #'prevalence incidence remission excess-mortality duration mortality relative-risk'.split())
# save results (do this last, because it removes things from the disease model that plotting function, etc, might need
keys = dismod3.utils.gbd_keys(region_list=[region], year_list=[year], sex_list=[sex])
dm.save('dm-%d-posterior-%s-%s-%s.json' % (dm.id, region, sex, year), keys_to_save=keys)
return dm
sys.path += ['..']
import pylab as pl
import pymc as mc
import dismod3
import book_graphics
reload(book_graphics)
results = {}
models = {}
for ii in range(2):
### @export 'load model'
dm = dismod3.load_disease_model(16391)
### @export 'set expert priors'
dm.params['global_priors']['level_value']['incidence'] = dict(value=0., age_before=1., age_after=99)
dm.params['global_priors']['smoothness']['prevalence']['amount'] = 'Slightly'
dm.params['global_priors']['heterogeneity']['prevalence'] = 'Slightly'
dm.params['global_priors']['level_value']['prevalence'] = dict(value=0., age_before=0, age_after=100)
dm.params['global_priors']['level_bounds']['prevalence'] = dict(lower=0., upper =.05)
dm.params['global_priors']['increasing']['prevalence'] = dict(age_start=0, age_end=0)
dm.params['global_priors']['decreasing']['prevalence'] = dict(age_start=100, age_end=100)
dm.params['sex_effect_prevalence'] = dict(mean=1, upper_ci=1.0001, lower_ci=.9999)
dm.params['time_effect_prevalence'] = dict(mean=1, upper_ci=1.0001, lower_ci=.9999)
dm.params['region_effect_prevalence'] = dict(std=.0001)
dm.params['covariates']['Study_level']['bias']['rate']['value'] = 0
for cv in dm.params['covariates']['Country_level']:
model.mu['abs_err'].mean(),
pl.median(pl.absolute(
model.mu['rel_err'].dropna())), model.mu['covered?'].mean())
print
data_simulation.initialize_results(model)
data_simulation.add_to_results(model, 'mu')
data_simulation.finalize_results(model)
print model.results
return model
if __name__ == '__main__':
region, sex, year = 'north_america_high_income', 'male', 1990
import fit_posterior, upload_fits
import data
import simplejson as json
## load the model from disk or from web
dm = dismod3.load_disease_model(24842)
dm.model = data.ModelData.from_gbd_jsons(json.loads(dm.to_json()))
data = upload_fits.merge_data_csvs(24842)
dm.model.input_data['mu_pred'] = data['mu_pred']
simulate_data(dm, region, sex, year)
fit_simulated(dm, region, sex, year)
store_results(dm, region, sex, year)
