These kind of model you usually need to rewrite them into a marginalized mixture model to be able to do effective influence. You can take a look at the following related topics:
Hello,
I’m new to PyMC3 and for getting started I tried to implement a Naive Bayes (NB) model as described in this blog post but for version 3.5 of PyMC. Here’s my initial attempt:
import numpy as np
import pymc3 as pm
K = 3 # number of topics
V = 4 # size of vocabulary
# 13 documents with 5 words each
data = np.array([[0, 0, 0, 0, 0],
[0, 0, 0, 0, 1],
[0, 0, 0, 1, 1],
[0, 0, 0, 0, 1],
[0, 0, 0, 0, 1],
…
[model1] ![model3|611x71]
[image]
[image]
Hi all,
I am new to PyMC3. I am having trouble implementing the occupancy model shown in the images above. The alpha and beta each have a prior distribution that follows a multivariate normal distribution.
with pm.Model() as _sampler:
beta = pm.Normal('beta', mu=0, sd=1000, shape=len(HYPERS['b_mu']))
alpha = pm.Normal('alpha', mu=0, sd=1000, shape=len(HYPERS['a_mu']))
…