Bayesian Networks using pymc5

Questions v5
,
1

I’m trying to solve a problem similar to this: Bayes Nets, Belief Networks, and PyMC in pymc5.
In the solution to the problem (Bayes Nets, Belief Networks, and PyMC - #2 by junpenglao), theono.shared is used to create a lookup table. To recreate this using pymc5, I think I need to convert theono.shared to pymc.Data but I’m not sure how to go about it. Any insight will be helpful, thanks!

1 Like
2

I am facing a similar issue as well, any help is appreciated.

3

You can use pytensor.shared just the same as before with theano. pm.Data is just a convenience around shared variables so they show in InferenceData and graphviz.

4

Does anyone make this wiork?
Thanks,

5

Hey I translated this to pymc5 a while ago. You can see it here.

1 Like