def __main__():
tobit_data = prepare_tobit_data()
ad_matrix = get_students_adjacency(tobit_data)
# here, try any of the models defined before.
fit, model = scaled_spare_car(tobit_data, ad_matrix)
# y_cens look ok though
# tau quite large -> 23, close to the largest „friend_group“
# larger phis now :) in negative and positive!
# investigate: can I use the normal_l(c)cdf function?
# fit, model = tobit_simple_model(tobit_data, scaled=True)
# fit, model = tobit_cum_sum_scaled(tobit_data)
# fit, model = tobit_vec_QR(tobit_data)
# note: this yields a expected values for β, but throws warnings for:
# - Rhat (though everything is 1)
# -
az.plot_trace(fit, compact=True)
az.plot_pair(fit, ['tau', 'alpha', 'sigma'], divergences=True)
# seems like I'm having a lot of divergences where:
# - sigma below 0.0025
# - alpha > 0.99 (would imply IAR)
# -> constraining helped a bit. But having region _around_ sigma = 0.08 and 0.04
az.plot_energy(fit)
def create_diagnostic_plots(idx,pdf_filename,fit,diag_pars,niter,nchain):
# Converting the Stan FIT object to Arviz InfereceData
samples = fit.extract(permuted=True) # Extracting parameter samples
data = az.from_pystan(fit)
tmp = data.posterior
var_names = list(tmp.data_vars)
# Filtering the list of parameters to plot
unwanted = {'losvd','spec','conv_spec','poly','bestfit','losvd_','losvd_mod','spec_pred','log_likelihood'}
vars_main = [e for e in var_names if e not in unwanted]
# Reading diagnostic parameters
accept_stat, stepsize, treedepth = np.zeros((niter,nchain)), np.zeros((niter,nchain)) , np.zeros((niter,nchain))
n_leapfrog, divergent, energy = np.zeros((niter,nchain)), np.zeros((niter,nchain)) , np.zeros((niter,nchain))
for j in range(nchain):
accept_stat[:,j] = diag_pars[j]['accept_stat__']
stepsize[:,j] = diag_pars[j]['stepsize__']
treedepth[:,j] = diag_pars[j]['treedepth__']
n_leapfrog[:,j] = diag_pars[j]['n_leapfrog__']
divergent[:,j] = diag_pars[j]['divergent__']
energy[:,j] = diag_pars[j]['energy__']
# Creating the plot in multiple PDF papges
pdf_pages = PdfPages(pdf_filename)
print(" - Sampler params")
plot_sampler_params(idx,accept_stat,stepsize,treedepth,n_leapfrog,divergent,energy)
pdf_pages.savefig()
print(" - Chains")
plot_chains(samples,vars_main)
pdf_pages.savefig()
# print(" - Trace plot [Main params]")
# az.plot_trace(data, var_names=vars_main)
# pdf_pages.savefig()
# print(" - Trace plot [LOSVD]")
# az.plot_trace(data, var_names=['losvd'])
# pdf_pages.savefig()
print(" - Pair plot")
az.plot_pair(data, var_names=vars_main, divergences=True, kind='kde', fill_last=False)
pdf_pages.savefig()
print(" - Autocorr plot")
az.plot_autocorr(data, var_names=vars_main)
pdf_pages.savefig()
print(" - Energy plot")
az.plot_energy(data)
pdf_pages.savefig()
pdf_pages.close()
return
def tobit_ifelse_model(tobit_data: pd.DataFrame):
"""
Use a loop instead of two matrices as preparation for using the adjacency matrix:
"""
censored_dict = get_datadict(tobit_data)
censored_loop_model = StanModel(
file=Path('models/tobit_students_ifelse.stan').open(),
extra_compile_args=["-w"])
censored_loop_fit = censored_loop_model.sampling(data=censored_dict,
iter=2000,
chains=4,
warmup=500)
az.plot_trace(censored_loop_fit)
az.plot_energy(censored_loop_fit)
cens_loop_res = censored_loop_fit.extract()
print('α: {}'.format(cens_loop_res['alpha'][501:].mean()))
print('β: {}'.format(cens_loop_res['beta'][501:].mean(axis=0)))
# yay works. intercept: 208.6, read: 2.70, math: 5.93, gen: -12.75, voc: -46.6
return censored_loop_fit, censored_loop_model
"""
Energy Plot
===========
_thumb: .7, .5
_example_title: Plot energy
"""
import matplotlib.pyplot as plt
import arviz as az
az.style.use("arviz-darkgrid")
data = az.load_arviz_data("centered_eight")
az.plot_energy(data, figsize=(12, 8))
plt.show()
'away_score' : away_score,
'npredict' : npredict,
'home_team_new' : home_team_new,
'away_team_new' : away_team_new
}
mod = pystan.StanModel(model_code=model_2)
fit = mod.sampling(data = data, iter=15000, warmup=1000, chains=4, control=dict(max_treedepth = 14, adapt_delta = .9))
"""### **ASSESSING MODEL**"""
print(fit)
az.style.use('arviz-darkgrid')
inf_data = az.convert_to_inference_data(fit)
az.plot_energy(inf_data)
az.plot_trace(fit,var_names=['home_score_new', 'away_score_new'])
az.plot_trace(fit, var_names=['att', 'def'], combined=True)
plt.style.use('ggplot')
_, ax = plt.subplots(1, 2, figsize=(15, 6))
az.plot_forest(fit, var_names="att",combined=True, ax=ax[0], kind='ridgeplot', ridgeplot_alpha=.5, ridgeplot_overlap=1.5, hdi_prob=.999, linewidth=.5)
ax[0].set_yticklabels(sorted(names, reverse=True))
ax[0].set_title('Estimated Attack Effect (Positive is Better)', loc='left')
ax[0].grid(True)
az.plot_forest(fit, var_names="def", combined=True, ax=ax[1], kind='ridgeplot', ridgeplot_alpha=.5, ridgeplot_overlap=1.5, colors='#99c2ff', hdi_prob=.999,
linewidth=.5)
ax[1].set_yticklabels(sorted(names, reverse=True))
ax[1].set_title('Estimated Defense Effect (Negative is Better)', loc='left')
ax[1].grid(True)
trace = pm.sample(2000,
tune=1000,
random_seed=RANDOM_SEED,
target_accept=0.95,
init="advi+adapt_diag")
with ic_model:
ppc = pm.sample_posterior_predictive(trace)
az_data = az.from_pymc3(trace=trace,
posterior_predictive=ppc,
model=ic_model,
coords={"episodes": epi_enc.classes_},
dims={"κ": ["episodes"]})
az.plot_energy(az_data, figsize=(6, 4))
vnames = ["κ", "β", "σ_y"]
sum_df = az.summary(az_data, var_names=vnames, round_to=3)
sum_df.index = ["κ: " + x for x in list(epi_enc.classes_)] + ["β", "σ_y"]
sum_df
# ## Estimates
#
# In the Table above, we present the results of the exponential decay model.
# The $\kappa$ parameter ranges from a low of 0.02 during the Great Recession (GFC)
# to a high of 0.4 in the aftermath of Katrina. The estimate for the current COVID
# episode is in between the 2 but subject to considerable uncertainty given we
# are still early in the process. As more data becomes available,
# the 95% credibility intervals will shrink but we have to be
# cognizant of the fact that any projections will of necessity be
"""
Energy Plot
===========
_thumb: .7, .5
"""
import arviz as az
data = az.load_arviz_data("centered_eight")
ax = az.plot_energy(data, figsize=(12, 8), backend="bokeh")
trace = pm.sample(3000, cores=2, return_inferencedata=True)
az.summary(trace)
az.summary(trace.posterior['intercept'])
pm.traceplot(trace)
plt.show()
az.plot_trace(trace)
az.plot_trace(trace, var_names=['intercept'])
plt.show()
az.plot_trace(trace, var_names=['sigma'])
plt.show()
az.plot_forest(trace, r_hat=True)
plt.show()
az.plot_posterior(trace)
plt.show()
az.plot_energy(trace)
plt.show()
az.plot_autocorr(trace, var_names=["intercept", "sigma"])
plt.show()
# map() by hand
def negPost(x, *args):
y = args[0]
p = -1 * sum(norm.logpdf(y, x[0], x[1]))[0] + norm.logpdf(
x[0], 178, 20) + uniform.logpdf(x[1], 0, 50)
return p
negPost((178, 3), d2)
secondDerivPost = nd.Derivative(negPost, n=2, full_output=True)
