This is PyMC3 version of http://nbviewer.ipython.org/github/tokyo-scipy/archive/blob/master/006/shima__shima/NaiveBayes.ipynb

pymc3_tutor2.py

import pymc as pm
import numpy as np

ndims = 2
nobs = 20

n = 1000
y_sample = np.random.binomial(1, 0.5, size=(n,))
x_sample=np.empty(n)
x_sample[y_sample==0] = np.random.normal(-1, 1, size=(n, ))[y_sample==0]
x_sample[y_sample==1] = np.random.normal(1, 1, size=(n, ))[y_sample==1]

with pm.Model() as model:
	p = pm.Beta('p', alpha=1.0, beta=1.0)
	y = pm.Bernoulli('y', p=p, observed=y_sample)
	mu0 = pm.Normal('mu0', mu=0., sd=1.)
	mu1 = pm.Normal('mu1', mu=0., sd=1.)

	mu = pm.Deterministic('mu', mu0 * (1-y_sample) + mu1 * y_sample)
	
	x = pm.Normal('x', mu=mu, sd=1., observed=x_sample)

with model:
	start = pm.find_MAP()
	step = pm.NUTS()
	trace = pm.sample(10000, step, start)

pm.traceplot(trace).savefig("result2.jpg")