Trying to run a glm: pm.glm.glm or pm.glm.GLM?

Questions
1

I am just getting started learning my way around pymc3 and am working through some of the examples.

One doc uses glm.glm
https://pymc-devs.github.io/pymc3/notebooks/GLM-robust.html

and another glm.GLM.from_formula
https://pymc-devs.github.io/pymc3/notebooks/getting_started.html#Generalized-Linear-Models

Any help would be appreciated. Like I said, just trying to learn

2

Hi @sthomas522

The correct use case should be glm.GLM.from_formula. For example in the GLM-robust the correct way is pm.glm.GLM.from_formula('y ~ x', data)

We are currently rebuilding our docs website, sorry about the confusion.

3

Thanks for the reply and clarification. This is a great package.

4

Also check out Bambi which is much more powerful: https://github.com/bambinos/bambi

5

Looks great - thank you. I like the mixed effect model formulas which are very similar to R.