def mixture_of_real_and_int():
loc = ed.Normal(loc=0., scale=1., name="loc")
flip = ed.Bernoulli(probs=0.5, name="flip")
if tf.equal(flip, 1):
x = ed.Normal(loc=loc, scale=0.5, sample_shape=5, name="x")
else:
x = ed.Poisson(rate=tf.nn.softplus(loc), sample_shape=3, name="x")
return x
def model(home_id, away_id, score1_obs=None, score2_obs=None):
# priors
alpha = ed.Normal(loc=0.0, scale=1.0, name="alpha")
sd_att = ed.StudenT(df=3.0, loc=0.0, scale=2.5, name="sd_att")
sd_def = ed.StudenT(df=3.0, loc=0.0, scale=2.5, name="sd_def")
home = ed.Normal(loc=0.0, scale=1.0, name="home") # home advantage
nt = len(np.unique(home_id))
# team-specific model parameters
attack = ed.Normal(loc=0, scale=sd_att, sample_shape=nt, name="attack")
defend = ed.Normal(loc=0, scale=sd_def, sample_shape=nt, name="defend")
# likelihood
theta1 = tf.exp(alpha + home + attack[home_id] - defend[away_id])
theta2 = tf.exp(alpha + attack[away_id] - defend[home_id])
s1 = ed.Poisson(theta1, name="s1")
s2 = ed.Poisson(theta2, name="s2")
return s1, s2