def test_1latent(self):
model_code = """
data {
int<lower=0> N;
int<lower=0,upper=1> x[N];
}
parameters {
real<lower=0,upper=1> p;
}
model {
p ~ beta(1.0, 1.0);
for (n in 1:N)
x[n] ~ bernoulli(p);
}
"""
with self.test_session():
model = ed.StanModel(model_code=model_code)
data = {'N': 10, 'x': [0, 1, 0, 1, 0, 1, 0, 1, 1, 1]}
zs = {'p': np.array(0.5)}
_test(model, data, zs)
def test_1d():
model_code = """
data {
int<lower=0> N;
int<lower=0,upper=1> y[N];
}
parameters {
real<lower=0,upper=1> theta;
}
model {
theta ~ beta(0.5, 0.5); // Jeffreys' prior
for (n in 1:N)
y[n] ~ bernoulli(theta);
}
"""
pystan_model = pystan.StanModel(model_code=model_code)
ed_model = ed.StanModel(model=pystan_model)
data = {'N': 10, 'y': [0, 1, 0, 1, 0, 1, 0, 1, 1, 1]}
zs = np.array([[0.5]])
_test(ed_model, pystan_model, data, zs)
zs = np.array([[0.4], [0.2], [0.2351], [0.6213]])
_test(ed_model, pystan_model, data, zs)
data = np.array([0, 1, 0, 0, 0, 0, 0, 0, 0, 1])
npmodel = BetaBernoulli(data)
model_code = """
data {
int<lower=0> N;
int<lower=0,upper=1> y[N];
}
parameters {
real<lower=0,upper=1> theta;
}
model {
theta ~ beta(1.0, 1.0);
for (n in 1:N)
y[n] ~ bernoulli(theta);
}
"""
data = dict(N=10, y=[0, 1, 0, 0, 0, 0, 0, 0, 0, 1])
model = ed.StanModel(model_code=model_code, data=data)
print(npmodel._py_log_prob(np.array([[0.5]], dtype=np.float32)))
print(model._py_log_prob(np.array([[0.5]], dtype=np.float32)))
print()
print(npmodel._py_log_prob(np.array([[0.314]], dtype=np.float32)))
print(model._py_log_prob(np.array([[0.314]], dtype=np.float32)))
print()
print(npmodel._py_log_prob(np.array([[0.682]], dtype=np.float32)))
print(model._py_log_prob(np.array([[0.682]], dtype=np.float32)))
Prior: Beta
Likelihood: Bernoulli
Variational model
Likelihood: Mean-field Beta
"""
import edward as ed
from edward.models import Variational, Beta
model_code = """
data {
int<lower=0> N;
int<lower=0,upper=1> y[N];
}
parameters {
real<lower=0,upper=1> theta;
}
model {
theta ~ beta(1.0, 1.0);
for (n in 1:N)
y[n] ~ bernoulli(theta);
}
"""
ed.set_seed(42)
model = ed.StanModel(model_code=model_code)
variational = Variational()
variational.add(Beta())
data = ed.Data(dict(N=10, y=[0, 1, 0, 0, 0, 0, 0, 0, 0, 1]))
inference = ed.MFVI(model, variational, data)
inference.run(n_iter=10000)
