missing_truncated.py

import pymc as pm
import numpy as np
 
missing_fill_value = 16.172
data = np.array([None, None, None, 12, 17, 20])
#data = np.where(data == np.array(None), missing_fill_value, data)
 
#masked_values = np.ma.masked_equal(data, value=missing_fill_value)
masked_values = np.ma.masked_array(data, np.equal(data, None), fill_value=10)

mu = pm.Normal('mu', 10, 1e-6, value=10)
tau = pm.Exponential('tau', 0.01, value=10)
 
x = pm.TruncatedNormal('x', mu=mu, tau=tau, a=7, b=27, value=data, observed=True)
 
M = pm.MCMC(locals())
M.sample(2000, burn=1000)

mu.summary()