def main(args):
baseball_dataset = pd.read_csv(DATA_URL, "\t")
train, _, player_names = train_test_split(baseball_dataset)
at_bats, hits = train[:, 0], train[:, 1]
nuts_kernel = NUTS(conditioned_model, adapt_step_size=True)
logging.info("Original Dataset:")
logging.info(baseball_dataset)
# (1) Full Pooling Model
posterior_fully_pooled = MCMC(nuts_kernel, num_samples=args.num_samples, warmup_steps=args.warmup_steps) \
.run(fully_pooled, at_bats, hits)
logging.info("\nModel: Fully Pooled")
logging.info("===================")
logging.info("\nphi:")
logging.info(
summary(posterior_fully_pooled,
sites=["phi"],
player_names=player_names)["phi"])
posterior_predictive = TracePredictive(fully_pooled,
posterior_fully_pooled,
num_samples=args.num_samples)
sample_posterior_predictive(posterior_predictive, baseball_dataset)
evaluate_log_predictive_density(fully_pooled, posterior_fully_pooled,
baseball_dataset)
# (2) No Pooling Model
posterior_not_pooled = MCMC(nuts_kernel, num_samples=args.num_samples, warmup_steps=args.warmup_steps) \
.run(not_pooled, at_bats, hits)
logging.info("\nModel: Not Pooled")
logging.info("=================")
logging.info("\nphi:")
logging.info(
summary(posterior_not_pooled, sites=["phi"],
player_names=player_names)["phi"])
posterior_predictive = TracePredictive(not_pooled,
posterior_not_pooled,
num_samples=args.num_samples)
sample_posterior_predictive(posterior_predictive, baseball_dataset)
evaluate_log_predictive_density(not_pooled, posterior_not_pooled,
baseball_dataset)
# (3) Partially Pooled Model
posterior_partially_pooled = MCMC(nuts_kernel, num_samples=args.num_samples, warmup_steps=args.warmup_steps) \
.run(partially_pooled, at_bats, hits)
logging.info("\nModel: Partially Pooled")
logging.info("=======================")
logging.info("\nSigmoid(alpha):")
logging.info(
summary(posterior_partially_pooled,
sites=["alpha"],
player_names=player_names,
transforms={"alpha": lambda x: 1. /
(1 + np.exp(-x))})["alpha"])
posterior_predictive = TracePredictive(partially_pooled,
posterior_partially_pooled,
num_samples=args.num_samples)
sample_posterior_predictive(posterior_predictive, baseball_dataset)
evaluate_log_predictive_density(partially_pooled,
posterior_partially_pooled,
baseball_dataset)
def sampling_prediction(self, svi, x_train, y_train, x_test,
num_samples = 1000):
posterior = svi.run(x_train, y_train)
trace_pred = TracePredictive(self.wrapped_model, posterior,
num_samples = num_samples)
post_pred = trace_pred.run(x_test, None)
sites= ['prediction', 'obs']
marginal = get_marginal(post_pred, sites)
site_stats = {}
for i in range(marginal.shape[1]):
site_name = sites[i]
marginal_site = pd.DataFrame(marginal[:, i]).transpose()
site_stats[site_name] = marginal_site.apply(pd.Series.describe,
axis=1)[["mean", "std"]]
mu = site_stats["prediction"]
y_o = site_stats["obs"]
return mu["mean"], mu["std"], y_o["mean"], y_o["std"]
def evaluate_log_predictive_density(model, model_trace_posterior, baseball_dataset):
"""
Evaluate the log probability density of observing the unseen data (season hits)
given a model and empirical distribution over the parameters.
"""
_, test, player_names = train_test_split(baseball_dataset)
at_bats_season, hits_season = test[:, 0], test[:, 1]
test_eval = TracePredictive(conditioned_model,
model_trace_posterior,
num_samples=args.num_samples)
test_eval.run(model, at_bats_season, hits_season)
trace_log_pdf = []
for tr in test_eval.exec_traces:
trace_log_pdf.append(tr.log_prob_sum())
# Use LogSumExp trick to evaluate $log(1/num_samples \sum_i p(new_data | \theta^{i})) $,
# where $\theta^{i}$ are parameter samples from the model's posterior.
posterior_pred_density = log_sum_exp(torch.stack(trace_log_pdf)) - math.log(len(trace_log_pdf))
logging.info("\nLog posterior predictive density")
logging.info("---------------------------------")
logging.info("{:.4f}\n".format(posterior_pred_density))
def evaluate_log_predictive_density(model, model_trace_posterior, baseball_dataset):
"""
Evaluate the log probability density of observing the unseen data (season hits)
given a model and empirical distribution over the parameters.
"""
_, test, player_names = train_test_split(baseball_dataset)
at_bats_season, hits_season = test[:, 0], test[:, 1]
test_eval = TracePredictive(conditioned_model,
model_trace_posterior,
num_samples=args.num_samples)
test_eval.run(model, at_bats_season, hits_season)
trace_log_pdf = []
for tr in test_eval.exec_traces:
trace_log_pdf.append(tr.log_prob_sum())
# Use LogSumExp trick to evaluate $log(1/num_samples \sum_i p(new_data | \theta^{i})) $,
# where $\theta^{i}$ are parameter samples from the model's posterior.
posterior_pred_density = log_sum_exp(torch.stack(trace_log_pdf)) - math.log(len(trace_log_pdf))
logging.info("\nLog posterior predictive density")
logging.info("---------------------------------")
logging.info("{:.4f}\n".format(posterior_pred_density))
sites = ["a", "bA", "bR", "bAR", "sigma"]
for site, values in summary(posterior, sites).items():
print("Site: {}".format(site))
print(values, "\n")
def wrapped_model(x_data, y_data):
pyro.sample("prediction", dist.Delta(model(x_data, y_data)))
# posterior predictive distribution we can get samples from
trace_pred = TracePredictive(wrapped_model,
posterior,
num_samples=1000)
post_pred = trace_pred.run(x_data, None)
post_summary = summary(post_pred, sites= ['prediction', 'obs'])
mu = post_summary["prediction"]
y = post_summary["obs"]
print("sample y data:")
print(y.head(10))
predictions = pd.DataFrame({
"cont_africa": x_data[:, 0],
"rugged": x_data[:, 1],
"mu_mean": mu["mean"],
"mu_perc_5": mu["5%"],
"mu_perc_95": mu["95%"],
posterior = svi.run(log_gdp, is_cont_africa, ruggedness)
sites = ["a", "bA", "bR", "bAR", "sigma"]
for site, values in summary(posterior, sites).items():
print("Site: {}".format(site))
print(values, "\n")
def wrapped_model(is_cont_africa, ruggedness, log_gdp):
pyro.sample("prediction", Delta(model(is_cont_africa, ruggedness,
log_gdp)))
# posterior predictive distribution we can get samples from
trace_pred = TracePredictive(wrapped_model, posterior, num_samples=1000)
post_pred = trace_pred.run(is_cont_africa, ruggedness, None)
post_summary = summary(post_pred, sites=['prediction', 'obs'])
mu = post_summary["prediction"]
y = post_summary["obs"]
print("sample y data:")
print(y.head(10))
predictions = pd.DataFrame({
"cont_africa": x_data[:, 0],
"rugged": x_data[:, 1],
"mu_mean": mu["mean"],
"mu_perc_5": mu["5%"],
"mu_perc_95": mu["95%"],
"y_mean": y["mean"],
def main(args):
pyro.set_rng_seed(args.rng_seed)
baseball_dataset = pd.read_csv(DATA_URL, "\t")
train, _, player_names = train_test_split(baseball_dataset)
at_bats, hits = train[:, 0], train[:, 1]
logging.info("Original Dataset:")
logging.info(baseball_dataset)
num_predictive_samples = args.num_samples * args.num_chains
# (1) Full Pooling Model
nuts_kernel = NUTS(fully_pooled)
posterior_fully_pooled = MCMC(nuts_kernel,
num_samples=args.num_samples,
warmup_steps=args.warmup_steps,
num_chains=args.num_chains).run(at_bats, hits)
logging.info("\nModel: Fully Pooled")
logging.info("===================")
logging.info("\nphi:")
logging.info(summary(posterior_fully_pooled, sites=["phi"], player_names=player_names)["phi"])
posterior_predictive = TracePredictive(fully_pooled,
posterior_fully_pooled,
num_samples=num_predictive_samples)
sample_posterior_predictive(posterior_predictive, baseball_dataset)
evaluate_log_predictive_density(posterior_predictive, baseball_dataset)
# (2) No Pooling Model
nuts_kernel = NUTS(not_pooled)
posterior_not_pooled = MCMC(nuts_kernel,
num_samples=args.num_samples,
warmup_steps=args.warmup_steps,
num_chains=args.num_chains).run(at_bats, hits)
logging.info("\nModel: Not Pooled")
logging.info("=================")
logging.info("\nphi:")
logging.info(summary(posterior_not_pooled, sites=["phi"], player_names=player_names)["phi"])
posterior_predictive = TracePredictive(not_pooled,
posterior_not_pooled,
num_samples=num_predictive_samples)
sample_posterior_predictive(posterior_predictive, baseball_dataset)
evaluate_log_predictive_density(posterior_predictive, baseball_dataset)
# (3) Partially Pooled Model
# TODO: remove once htps://github.com/uber/pyro/issues/1458 is resolved
if "CI" not in os.environ:
nuts_kernel = NUTS(partially_pooled)
posterior_partially_pooled = MCMC(nuts_kernel,
num_samples=args.num_samples,
warmup_steps=args.warmup_steps,
num_chains=args.num_chains).run(at_bats, hits)
logging.info("\nModel: Partially Pooled")
logging.info("=======================")
logging.info("\nphi:")
logging.info(summary(posterior_partially_pooled, sites=["phi"],
player_names=player_names)["phi"])
posterior_predictive = TracePredictive(partially_pooled,
posterior_partially_pooled,
num_samples=num_predictive_samples)
sample_posterior_predictive(posterior_predictive, baseball_dataset)
evaluate_log_predictive_density(posterior_predictive, baseball_dataset)
# (4) Partially Pooled with Logit Model
nuts_kernel = NUTS(partially_pooled_with_logit)
posterior_partially_pooled_with_logit = MCMC(nuts_kernel,
num_samples=args.num_samples,
warmup_steps=args.warmup_steps,
num_chains=args.num_chains).run(at_bats, hits)
logging.info("\nModel: Partially Pooled with Logit")
logging.info("==================================")
logging.info("\nSigmoid(alpha):")
logging.info(summary(posterior_partially_pooled_with_logit,
sites=["alpha"],
player_names=player_names,
transforms={"alpha": lambda x: 1. / (1 + (-x).exp())})["alpha"])
posterior_predictive = TracePredictive(partially_pooled_with_logit,
posterior_partially_pooled_with_logit,
num_samples=num_predictive_samples)
sample_posterior_predictive(posterior_predictive, baseball_dataset)
evaluate_log_predictive_density(posterior_predictive, baseball_dataset)