results['Offset lognormal'] = dict(pi=pi.stats(), pred=pred.stats())
### @export 'save'
pi_median = []
pi_spread = []
for i, k in enumerate(results):
pi_median.append(results[k]['pi']['quantiles'][50])
pi_spread.append(results[k]['pi']['95% HPD interval'][1]-results[k]['pi']['95% HPD interval'][0])
min_est = min(pi_median).round(4)
max_est = max(pi_median).round(4)
min_spread = min(pi_spread).round(4)
max_spread = max(pi_spread).round(4)
book_graphics.save_json('schiz_forest.json', vars())
### data only plot, for computational infrastructure appendix
book_graphics.forest_plot(r, n, data_labels=cy,
xmax=.0115,
subplot_params=dict(bottom=.1, right=.99, top=.95, left=.15),
figparams=book_graphics.quarter_page_params,
fname='book/graphics/ci-prev_meta_analysis-schiz_data.png')
### master graphic of data and models, for rate model section of stats chapter
book_graphics.forest_plot(r, n, data_labels=cy,
xmax=.0115,
model_keys=['Binomial', 'Poisson', 'Beta binomial', 'Negative binomial', 'Normal', 'Lognormal', 'Offset lognormal'],
results=results,
#subplot_params=dict(bottom=.1, right=.99, top=.95, left=.15),
linewidth=1,
label='Uncertainty interval')
pl.errorbar(data[:, 0],
data[:, 1],
yerr=data[:, 2] * 1.96,
fmt='gs',
mec='white',
mew=0,
ms=10)
pl.axis([0, 100, 0, 10])
pl.text(10,
9,
'%s ($\\rho = %d$)' % (smoothness, scale[smoothness]),
ha='left',
va='top')
if col == 0:
pl.ylabel('Rate (per 1000 PY)')
else:
pl.yticks([])
pl.xticks([25, 50, 75])
pl.xlabel('Age (Years)')
pl.subplots_adjust(left=.1, bottom=.2, top=.95, right=.95, wspace=0)
pl.savefig('smoothness_priors.png')
pl.savefig('smoothness_priors.pdf')
### @export 'store-results'
book_graphics.save_json('age_patterns.json', vars())
dm.params['covariates']['Study_level']['bias']['rate']['value'] = 0
for cv in dm.params['covariates']['Country_level']:
dm.params['covariates']['Country_level'][cv]['rate']['value'] = 0
# TODO: set bounds on remission and excess-mortality in the second time through
### @export 'initialize model data'
region = 'north_america_high_income'
year = 1990
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 = {}
import fit_world
fit_world.fit_world(dm)
models[ii] = dm
results[ii] = dict(rate_stoch=dm.vars['prevalence+world+all+all']['rate_stoch'].stats(), dispersion=dm.vars['prevalence+world+all+all']['dispersion'].stats())
### @export 'save'
for d in dm.data:
d['sex'] = 'male' # otherwise tile plot shows it twice
dismod3.plotting.tile_plot_disease_model(dm, dismod3.utils.gbd_keys(['incidence', 'remission', 'excess-mortality', 'prevalence'], ['world'], ['all'], ['all']),
plot_prior=False, print_sample_size=False)
pl.show()
book_graphics.save_json('hep_c.json', vars())
# pl.plot([-100], [0], marker[j], label=label[j]) # HACK: off-screen point with label to appear in legend
for j in range(2):
vars = results[j].vars[dismod3.utils.gbd_key_for('prevalence', region,
year, sex)]
stats = vars['predicted_rates'].stats()
y = stats['quantiles'][50]
yerr = [y - stats['quantiles'][2.5], stats['quantiles'][97.5] - y]
pl.errorbar(sorted_indices + .25 * (j + 1),
y,
yerr,
mew=1,
ms=5,
mec='white',
fmt=marker[j],
label=label[j])
results['posterior_dispersion_%d' % j] = vars['dispersion'].stats()
results['posterior_beta_%d' % j] = vars['study_coeffs'].stats()
pl.xlabel('PMS prevalence data from systematic review')
pl.xticks([])
pl.yticks([0, .2, .4, .6, .8, 1])
pl.ylabel('Prevalence (per 1)')
pl.axis([-.5, 11.5, -.01, 1.])
pl.legend(loc='upper left', fancybox=True, shadow=True, numpoints=1)
pl.savefig('pms_ppc.pdf')
book_graphics.save_json('pms_ppc.json', vars())
zorder=1)
pl.plot(mesh, f_n,
'-', color='black', linewidth=1,
zorder=2)
#pl.plot(mesh, sm.f_eval.stats()['quantiles'][50],
# 'k-', linewidth=3, label='Median')
pl.plot(mesh, sm.f_eval.stats()['95% HPD interval'],
'k--', linewidth=1, label='Uncertainty interval')
pl.errorbar(data[:,0], data[:,1], yerr=data[:,2]*1.96,
fmt='gs',
mec='white', mew=0, ms=10)
pl.axis([0, 100, 0, 10])
pl.text(10, 9, '%s ($\\rho = %d$)' % (smoothness, scale[smoothness]), ha='left', va='top')
if col == 0:
pl.ylabel('Rate (per 1000 PY)')
else:
pl.yticks([])
pl.xticks([25,50,75])
pl.xlabel('Age (Years)')
pl.subplots_adjust(left=.1, bottom=.2, top=.95, right=.95, wspace=0)
pl.savefig('smoothness_priors.png')
pl.savefig('smoothness_priors.pdf')
### @export 'store-results'
book_graphics.save_json('age_patterns.json', vars())
results[grid] = dm
try:
results['dic_%s'%grid] = dm.mcmc.dic
except Exception, e:
print e
results['dic_%s'%grid] = 'TK'
pl.figure(**book_graphics.quarter_page_params)
for ii, grid in enumerate('abcd'):
dm = results[grid]
pl.subplot(1,4,ii+1)
dismod3.plotting.plot_intervals(dm, [d for d in dm.data if dm.relevant_to(d, 'prevalence', region, year, sex)],
color='black', print_sample_size=False, alpha=1., plot_error_bars=False,
linewidth=2)
book_graphics.plot_rate(dm, dismod3.utils.gbd_key_for('prevalence', region, year, sex), linestyle='-')
pl.axis([10, 60, -.05, .8])
pl.xlabel('Age (Years)')
pl.xticks([15,35,55])
if ii == 0:
pl.yticks([0, .2, .4, .6], [0, 20, 40, 60])
pl.ylabel('Prevalence (per 100)')
else:
pl.yticks([0, .2, .4, .6], ['', '', '', ''])
pl.text(12, .75, '(%s)'% grid, va='top', ha='left')
pl.subplots_adjust(wspace=0, bottom=.15, left=.1, right=.99, top=.97)
pl.savefig('pms_grids.pdf')
book_graphics.save_json('pms_grids.json', vars())
### @export 'negative-binomial_dispersion-alt_prior'
pi = mc.Uniform('pi', lower=0, upper=1, value=.5)
log_10_delta = mc.Normal('log_10_delta', mu=mu_log_10_delta, tau=.25**-2)
delta = mc.Lambda('delta', lambda x=log_10_delta: 5 + 10**x)
@mc.potential
def obs(pi=pi, delta=delta):
return mc.negative_binomial_like(r * n, pi * n, delta)
@mc.deterministic
def pred(pi=pi, delta=delta):
return mc.rnegative_binomial(pi * n_pred, delta) / float(n_pred)
### @export 'negative-binomial_dispersion-fit_alt_prior'
mc.MCMC([pi, log_10_delta, delta, obs, pred]).sample(iter,
burn,
thin,
verbose=False,
progress_bar=False)
key = 'Neg. Binom., $\mu_{\log\delta}=%d$' % mu_log_10_delta
results[key] = dict(pred=pred.stats(), pi=pi.stats())
model_keys.append(key)
#mc.Matplot.plot(pi)
#mc.Matplot.plot(delta)
book_graphics.save_json('poisson_model.json', vars())
book_graphics.forest_plot(fname='neg_binom_priors.pdf', **vars())
pl.show()
model_keys = ['Poisson', 'Negative Binomial']
### @export 'negative-binomial_dispersion-prior-exploration'
for mu_log_10_delta in [1,2,3]:
### @export 'negative-binomial_dispersion-alt_prior'
pi = mc.Uniform('pi', lower=0, upper=1, value=.5)
log_10_delta = mc.Normal('log_10_delta', mu=mu_log_10_delta, tau=.25**-2)
delta = mc.Lambda('delta', lambda x=log_10_delta: 5 + 10**x)
@mc.potential
def obs(pi=pi, delta=delta):
return mc.negative_binomial_like(r*n, pi*n, delta)
@mc.deterministic
def pred(pi=pi, delta=delta):
return mc.rnegative_binomial(pi*n_pred, delta) / float(n_pred)
### @export 'negative-binomial_dispersion-fit_alt_prior'
mc.MCMC([pi, log_10_delta, delta, obs, pred]).sample(iter, burn, thin, verbose=False, progress_bar=False)
key = 'Neg. Binom., $\mu_{\log\delta}=%d$'%mu_log_10_delta
results[key] = dict(pred=pred.stats(), pi=pi.stats())
model_keys.append(key)
#mc.Matplot.plot(pi)
#mc.Matplot.plot(delta)
book_graphics.save_json('poisson_model.json', vars())
book_graphics.forest_plot(fname='neg_binom_priors.pdf', **vars())
pl.show()
# print e
# pl.plot(i+.25*(j+1), vars['predicted_rates'].value[i], marker[j], label=label[j])
# results['posterior_dispersion_%d'%j] = 'TK'
#for j in range(2):
# pl.plot([-100], [0], marker[j], label=label[j]) # HACK: off-screen point with label to appear in legend
for j in range(2):
vars = results[j].vars[dismod3.utils.gbd_key_for('prevalence', region, year, sex)]
stats = vars['predicted_rates'].stats()
y = stats['quantiles'][50]
yerr = [y-stats['quantiles'][2.5], stats['quantiles'][97.5]-y]
pl.errorbar(sorted_indices+.25*(j+1), y, yerr,
mew=1, ms=5, mec='white',
fmt=marker[j], label=label[j])
results['posterior_dispersion_%d'%j] = vars['dispersion'].stats()
results['posterior_beta_%d'%j] = vars['study_coeffs'].stats()
pl.xlabel('PMS prevalence data from systematic review')
pl.xticks([])
pl.yticks([0, .2, .4, .6, .8, 1])
pl.ylabel('Prevalence (per 1)')
pl.axis([-.5, 11.5,-.01,1.])
pl.legend(loc='upper left', fancybox=True, shadow=True, numpoints=1)
pl.savefig('pms_ppc.pdf')
book_graphics.save_json('pms_ppc.json', vars())
mec='k',
mew=1,
label='%d Observation' % v)
else:
pl.plot([-100], [1],
'o',
color='none',
ms=pl.sqrt(v) * 5 + 2,
mec='k',
mew=1,
label='%d Observations' % v)
pl.legend(loc='lower left', fancybox=True, shadow=True, numpoints=1)
pl.xticks()
pl.yticks()
pl.xlabel('Mean of age group (years)')
pl.ylabel('Width of age group (years)')
pl.axis([-5, 110., .6, 500.])
pl.subplots_adjust(left=.1, right=.99, bottom=.15, top=.95)
pl.savefig('book/graphics/af_age_groups_scatter.pdf')
### @export 'save-results'
book_graphics.save_json('af_age_groups.json', {
'most_freq_cnt': most_freq_cnt,
'rows_total': rows_total
})
pl.show()
mc.MCMC([pi, obs, pred]).sample(20000, 10000, 10, verbose=False, progress_bar=False)
pl.figure(**book_graphics.quarter_page_params)
sorted_indices = r.argsort().argsort()
jitter = mc.rnormal(0, 0.1 ** -2, len(pred.trace()))
for i, s_i in enumerate(sorted_indices):
if i == 0:
label = "Predicted distribution"
else:
label = "_nolabel_"
pl.plot(s_i + jitter, pred.trace()[:, i] / float(n[i]), "ko", mew=0, alpha=0.25, zorder=-100, label=label)
pl.errorbar(
sorted_indices, r, yerr=1.96 * pl.sqrt(r * (1 - r) / n), fmt="ks", mew=1, ms=5, mec="white", label="Observed value"
)
pl.xticks([])
pl.yticks([0, 0.002, 0.004, 0.006, 0.008, 0.01])
pl.ylabel("Rate ($r$)")
pl.axis([-0.5, 15.5, -0.0001, 0.0121])
pl.legend(loc="upper center", numpoints=1, fancybox=True, shadow=True)
pl.savefig("book/graphics/binomial-model-ppc.pdf")
pl.savefig("book/graphics/binomial-model-ppc.png")
### @export 'save-vars'
book_graphics.save_json("binomial_model.json", vars())
pl.show()
pl.figure(**book_graphics.quarter_page_params)
for ii, grid in enumerate('abcd'):
dm = results[grid]
pl.subplot(1, 4, ii + 1)
dismod3.plotting.plot_intervals(dm, [
d
for d in dm.data if dm.relevant_to(d, 'prevalence', region, year, sex)
],
color='black',
print_sample_size=False,
alpha=1.,
plot_error_bars=False,
linewidth=2)
book_graphics.plot_rate(dm,
dismod3.utils.gbd_key_for('prevalence', region,
year, sex),
linestyle='-')
pl.axis([10, 60, -.05, .8])
pl.xlabel('Age (Years)')
pl.xticks([15, 35, 55])
if ii == 0:
pl.yticks([0, .2, .4, .6], [0, 20, 40, 60])
pl.ylabel('Prevalence (per 100)')
else:
pl.yticks([0, .2, .4, .6], ['', '', '', ''])
pl.text(12, .75, '(%s)' % grid, va='top', ha='left')
pl.subplots_adjust(wspace=0, bottom=.15, left=.1, right=.99, top=.97)
pl.savefig('pms_grids.pdf')
book_graphics.save_json('pms_grids.json', vars())
beta = mc.Uninformative('beta', value=10 * (1 - pop_A_prev))
pi = mc.Beta('pi', alpha, beta, value=[pop_A_prev, pop_B_prev, .02])
@mc.potential
def obs(pi=pi):
return pop_A_prev*pop_A_N*pl.log(pi[0]) + (1-pop_A_prev)*pop_A_N*pl.log(1-pi[0]) \
+ pop_B_prev*pop_B_N*pl.log(pi[1]) + (1-pop_B_prev)*pop_B_N*pl.log(1-pi[1])
pop_C_N = 50000
pop_C_k = mc.Binomial('pop_C_k', pop_C_N, pi[2])
mc.MCMC([alpha, beta, pi, obs, pop_C_k]).sample(200000,
100000,
20,
verbose=False,
progress_bar=False)
pop_C_prev = pop_C_k.stats()['quantiles'][50] / float(pop_C_N)
pop_C_prev_per_1000 = '%.0f' % (pop_C_prev * 1000)
print pop_C_prev_per_1000
pop_C_ui = pop_C_k.stats()['95% HPD interval'] / float(pop_C_N)
pop_C_ui_per_1000 = '[%.0f, %.0f]' % tuple(pop_C_ui * 1000)
print pop_C_ui_per_1000
### @export 'save-vars'
book_graphics.save_json('beta_binomial_model.json', vars())
pl.show()
mc.MCMC([pi, zeta, sigma, obs, pred]).sample(iter, burn, thin)
results['Offset lognormal'] = dict(pi=pi.stats(), pred=pred.stats())
### @export 'save'
pi_median = []
pi_spread = []
for i, k in enumerate(results):
pi_median.append(results[k]['pi']['quantiles'][50])
pi_spread.append(results[k]['pi']['95% HPD interval'][1] -
results[k]['pi']['95% HPD interval'][0])
min_est = min(pi_median).round(4)
max_est = max(pi_median).round(4)
min_spread = min(pi_spread).round(4)
max_spread = max(pi_spread).round(4)
book_graphics.save_json('zero_forest.json', vars())
### master graphic of data and models, for zeros subsection of rate models section of stats chapter
book_graphics.forest_plot(r,
n,
xmax=.0005,
model_keys=[
'Zero-inflated beta binomial', 'Beta binomial',
'Negative binomial', 'Lognormal',
'Offset lognormal'
],
results=results,
pi_true=pi_true,
fname='zero_forest.pdf')
# set expert priors and other model parameters
dm.params['global_priors']['level_value']['incidence']['age_before'] = 10
dm.params['global_priors']['level_value']['incidence']['age_after'] = 99
#dm.params['global_priors']['smoothness']['incidence']['age_start'] = 10
dm.params['global_priors']['level_bounds']['remission']['upper'] = .05
dm.params['global_priors']['level_value']['prevalence']['age_before'] = 10
dm.params['global_priors']['smoothness']['relative_risk']['amount'] = 'Very'
dm.params['covariates']['Country_level']['LDI_id_Updated_7July2011']['rate'][
'value'] = 0
dm.params['covariates']['Study_level']['cv_past_year']['rate']['value'] = 1
# clear any fit and priors
dm.clear_fit()
dm.clear_empirical_prior()
dismod3.neg_binom_model.covariate_hash = {}
# initialize model data
#dismod3.neg_binom_model.fit_emp_prior(dm, 'prevalence')
import fit_world
fit_world.fit_world(dm)
import fit_posterior
fit_posterior.fit_posterior(dm, 'north_america_high_income', 'female', '2005')
### @export 'save'
book_graphics.save_json('bipolar.json', vars())
### @export 'save'
for d in dm.data:
d['sex'] = 'male' # otherwise tile plot shows it twice
reload(book_graphics)
book_graphics.plot_age_patterns(
dm,
region='world',
year='all',
sex='all',
rate_types='remission incidence prevalence'.split())
pl.subplot(1, 3, 1)
pl.yticks([0, .04, .08], [0, 4, 8])
pl.ylabel('Rate (Per 100 PY)')
#pl.subplot(1,4,2)
#pl.yticks([0, .04, .08], [0,4,8])
pl.subplot(1, 3, 2)
pl.yticks([0, .0025, .0051], [0, .25, .5])
pl.subplot(1, 3, 3)
pl.yticks([0, .025, .05], [0, 2.5, 5])
pl.savefig('hep_c-consistent%d.pdf' % ii)
pl.show()
book_graphics.save_json('hep_c.json', results)
@mc.deterministic
def pred(pi=pi, delta=delta):
return mc.rnegative_binomial(pi*n_pred, delta) / float(n_pred)
## fit model
mc.MCMC([pi, delta, obs, pred]).sample(iter, burn, thin)
## record results
for i, stoch in enumerate([pi, pred]):
median = stoch.stats()['quantiles'][50]
residuals[i].append(pi_true - median)
lb, ub = stoch.stats()['95% HPD interval']
coverage[i].append(lb <= pi_true <= ub)
### @export 'summarize-results'
bias = {}
rmse = {}
percent_coverage = {}
for i, s in enumerate(['pi', 'pred']):
bias[s] = '%.5f' % pl.mean(residuals[i])
rmse[s] = '%.3f' % pl.rms_flat(residuals[i])
percent_coverage[s] = '%.2f' % pl.mean(coverage[i])
print 'bias', bias
print 'rmse', rmse
print 'percent coverage', percent_coverage
book_graphics.save_json('neg_binom_sim.json', vars())
# set expert priors and other model parameters
dm.params['global_priors']['level_value']['incidence']['age_before'] = 10
dm.params['global_priors']['level_value']['incidence']['age_after'] = 99
#dm.params['global_priors']['smoothness']['incidence']['age_start'] = 10
dm.params['global_priors']['level_bounds']['remission']['upper'] = .05
dm.params['global_priors']['level_value']['prevalence']['age_before'] = 10
dm.params['global_priors']['smoothness']['relative_risk']['amount'] = 'Very'
dm.params['covariates']['Country_level']['LDI_id_Updated_7July2011']['rate']['value'] = 0
dm.params['covariates']['Study_level']['cv_past_year']['rate']['value'] = 1
# clear any fit and priors
dm.clear_fit()
dm.clear_empirical_prior()
dismod3.neg_binom_model.covariate_hash = {}
# initialize model data
#dismod3.neg_binom_model.fit_emp_prior(dm, 'prevalence')
import fit_world
fit_world.fit_world(dm)
import fit_posterior
fit_posterior.fit_posterior(dm, 'north_america_high_income', 'female', '2005')
### @export 'save'
book_graphics.save_json('bipolar.json', vars())
linestyle = dict(slightly='solid', very='dashed', moderately='dotted')
for col, smoothness in enumerate(['slightly', 'moderately', 'very']):
mesh = pl.arange(-101, 101, .1)
C = mc.gp.FullRankCovariance(mc.gp.matern.euclidean,
amp=1.,
scale=rho[smoothness],
diff_degree=2)
pl.plot(mesh, C(mesh, [0]),
marker='', color='black', linewidth=1,
linestyle=linestyle[smoothness],
zorder=1,
label='%s ($\\rho = %d$)' % (smoothness.capitalize(), rho[smoothness]))
pl.xticks([-25, 0, 25,50,75])
pl.xlabel('$\\Delta$ Age (Years)')
pl.yticks([0, .25, .5, .75, 1.0])
pl.ylabel('Autocovariance')
pl.axis([-30, 90, -.1, 1.1])
pl.legend(loc='upper right', fancybox=True, shadow=True)
pl.subplots_adjust(left=.1, bottom=.2, top=.95, right=.95, wspace=0)
pl.savefig('smoothness_covariance.png')
pl.savefig('smoothness_covariance.pdf')
### @export 'store-results'
book_graphics.save_json('age_pattern_covariance.json', vars())
C = mc.gp.FullRankCovariance(mc.gp.matern.euclidean,
amp=1.,
scale=rho[smoothness],
diff_degree=2)
pl.plot(mesh,
C(mesh, [0]),
marker='',
color='black',
linewidth=1,
linestyle=linestyle[smoothness],
zorder=1,
label='%s ($\\rho = %d$)' %
(smoothness.capitalize(), rho[smoothness]))
pl.xticks([-25, 0, 25, 50, 75])
pl.xlabel('$\\Delta$ Age (Years)')
pl.yticks([0, .25, .5, .75, 1.0])
pl.ylabel('Autocovariance')
pl.axis([-30, 90, -.1, 1.1])
pl.legend(loc='upper right', fancybox=True, shadow=True)
pl.subplots_adjust(left=.1, bottom=.2, top=.95, right=.95, wspace=0)
pl.savefig('smoothness_covariance.png')
pl.savefig('smoothness_covariance.pdf')
### @export 'store-results'
book_graphics.save_json('age_pattern_covariance.json', vars())
pop_A_N = 50000
pop_B_prev = .006
pop_B_N = 50000
alpha = mc.Uninformative('alpha', value=10*pop_A_prev)
beta = mc.Uninformative('beta', value=10*(1-pop_A_prev))
pi = mc.Beta('pi', alpha, beta, value=[pop_A_prev, pop_B_prev, .02])
@mc.potential
def obs(pi=pi):
return pop_A_prev*pop_A_N*pl.log(pi[0]) + (1-pop_A_prev)*pop_A_N*pl.log(1-pi[0]) \
+ pop_B_prev*pop_B_N*pl.log(pi[1]) + (1-pop_B_prev)*pop_B_N*pl.log(1-pi[1])
pop_C_N = 50000
pop_C_k = mc.Binomial('pop_C_k', pop_C_N, pi[2])
mc.MCMC([alpha, beta, pi, obs, pop_C_k]).sample(200000,100000,20, verbose=False, progress_bar=False)
pop_C_prev = pop_C_k.stats()['quantiles'][50] / float(pop_C_N)
pop_C_prev_per_1000 = '%.0f' % (pop_C_prev*1000)
print pop_C_prev_per_1000
pop_C_ui = pop_C_k.stats()['95% HPD interval'] / float(pop_C_N)
pop_C_ui_per_1000 = '[%.0f, %.0f]' % tuple(pop_C_ui*1000)
print pop_C_ui_per_1000
### @export 'save-vars'
book_graphics.save_json('beta_binomial_model.json', vars())
pl.show()
1./((s_pred/(pi+zeta))**2 + sigma**2))) \
- zeta
### @export 'offset-log-normal-fit-and-store'
mc.MCMC([pi, zeta, sigma, obs, pred]).sample(iter, burn, thin)
results['Offset lognormal'] = dict(pi=pi.stats(), pred=pred.stats())
### @export 'save'
pi_median = []
pi_spread = []
for i, k in enumerate(results):
pi_median.append(results[k]['pi']['quantiles'][50])
pi_spread.append(results[k]['pi']['95% HPD interval'][1]-results[k]['pi']['95% HPD interval'][0])
min_est = min(pi_median).round(4)
max_est = max(pi_median).round(4)
min_spread = min(pi_spread).round(4)
max_spread = max(pi_spread).round(4)
book_graphics.save_json('zero_forest.json', vars())
### master graphic of data and models, for zeros subsection of rate models section of stats chapter
book_graphics.forest_plot(r, n,
xmax=.0005,
model_keys=['Zero-inflated beta binomial', 'Beta binomial', 'Negative binomial', 'Lognormal', 'Offset lognormal'],
results=results,
pi_true=pi_true,
fname='zero_forest.pdf')
fit_world.fit_world(dm, generate_diagnostic_plots=False, store_results=False, map_only=False)
models[ii] = dm
#results[ii] = dict(rate_stoch=dm.vars['prevalence+world+all+all']['rate_stoch'].stats(), dispersion=dm.vars['prevalence+world+all+all']['dispersion'].stats())
### @export 'save'
for d in dm.data:
d['sex'] = 'male' # otherwise tile plot shows it twice
reload(book_graphics)
book_graphics.plot_age_patterns(dm, region='world', year='all', sex='all',
rate_types='remission incidence prevalence'.split())
pl.subplot(1,3,1)
pl.yticks([0, .04, .08], [0,4,8])
pl.ylabel('Rate (Per 100 PY)')
#pl.subplot(1,4,2)
#pl.yticks([0, .04, .08], [0,4,8])
pl.subplot(1,3,2)
pl.yticks([0, .0025, .0051], [0,.25,.5])
pl.subplot(1,3,3)
pl.yticks([0, .025, .05], [0,2.5,5])
pl.savefig('hep_c-consistent%d.pdf' % ii)
pl.show()
book_graphics.save_json('hep_c.json', results)
for a_0 in range(101):
for a_1 in range(101):
if hist[a_0, a_1] <= 0.:
continue
x_i = .5 * (a_0 + a_1)
y_i = a_1 - a_0 + 1 # epidemiologist notation age 0-4 is five years
pl.semilogy([x_i], [y_i], 'o', color='none', ms=pl.sqrt(hist[a_0,a_1])*5+2, mec='k', mew=1)
#pl.text(x_i, y_i, '%d'%hist[a_0, a_1], ha='center', va='center')
for v in [1, 5, 10]:
if v == 1: pl.plot([-100], [1], 'o', color='none', ms=pl.sqrt(v)*5+2, mec='k', mew=1, label='%d Observation'%v)
else: pl.plot([-100], [1], 'o', color='none', ms=pl.sqrt(v)*5+2, mec='k', mew=1, label='%d Observations'%v)
pl.legend(loc='lower left', fancybox=True, shadow=True, numpoints=1)
pl.xticks()
pl.yticks()
pl.xlabel('Mean of age group (years)')
pl.ylabel('Width of age group (years)')
pl.axis([-5, 110., .6, 500.])
pl.subplots_adjust(left=.1, right=.99, bottom=.15, top=.95)
pl.savefig('book/graphics/af_age_groups_scatter.pdf')
### @export 'save-results'
book_graphics.save_json('af_age_groups.json', {'most_freq_cnt': most_freq_cnt, 'rows_total': rows_total})
pl.show()