Why is posterior same as prior despite observing data?

Questions
3

Thanks, this works!

I was also able to use theano's ifelse:

from theano.ifelse import ifelse

with pm.Model() as model:
    weekend = pm.Bernoulli('weekend', 2/7)
    rate = ifelse(weekend, 3, 10)
    bus = pm.Poisson('buses', mu=rate, observed=4)
    trace = pm.sample(10000)

Indeed, the docs for Deterministic say that the var parameter should be a “theano variable”:

pymc3.model.Deterministic(name, var, model=None)¶
Create a named deterministic variable

Parameters
name str
var theano variables

But perhaps it should feature more prominently in the docs (even if this is self-evident for someone who understands how PyMC works under the hood).

1 Like