GLM logistic

[OriolAbril]

File: https://github.com/pymc-devs/pymc-examples/blob/main/examples/generalized_linear_models/GLM-logistic.ipynb
Reviewers:

The sections below may still be pending. If so, the issue is still available, it simply doesn't
have specific guidance yet. Please refer to this overview of updates

Known changes needed

Changes listed in this section should all be done at some point in order to get this
notebook to a "Best Practices" state. However, these are probably not enough!
Make sure to thoroughly review the notebook and search for other updates.

General updates

ArviZ related

Changes for discussion

Changes listed in this section are up for discussion, these are ideas on how to improve
the notebook but may not have a clear implementation, or fix some know issue only partially.

General updates

ArviZ related

Notes

Exotic dependencies

Computing requirements

[twiecki]

This is also using the deprecated glm submodule which we should remove and implement a logistic regression directly.

[chiral-carbon]

hey! can I take this up?

[twiecki]

Totally, thanks for offering to help.

…

On Thu, Jul 22, 2021, 19:00 Abhipsha Das ***@***.***> wrote: hey! can I take this up? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#82 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAFETGB7RDHPWFMT7HDS73LTZBFEJANCNFSM42DGPOHQ> .

[drbenvincent]

I can update this to v4

[drbenvincent]

@OriolAbril / @twiecki
I've started to port this, but noticed that it overlaps a lot with an official Bambi example (see https://bambinos.github.io/bambi/main/notebooks/model_comparison.html)

I will definitely link to that, but I'm wondering if this notebook is essentially redundant?

Let me know if we should kill this or just continue to port what is there.

[ghost]

I vote in favor of removing. We should take advantage of any opportunity to reduce the size of the example library.

[aloctavodia]

Agreed. Is there a place in PyMC were we redirect/point to bambi for GLMs?

[drbenvincent]

Agreed. Is there a place in PyMC were we redirect/point to bambi for GLMs?

Some of the GLM notebooks do hand coded PyMC models. Some include Bambi only. Some contain a mix.

If we were going to move towards being consistent, my preference would be to focus on hand coded PyMC models but with linking to Bambi and possibly including non-executed Bambi code in markdown cells.

Why? I think it makes sense to avoid creating a dependency upon the Bambi API. If it changes then we have to update the PyMC examples. This also then creates redundancy. It also goes against the whole 'lego block' approach of building your own models with PyMC.

I think Bambi is fantastic, but for the PyMC examples, my feeling is that we should avoid Bambi only notebooks.

But in the mean time I'll open a PR to remove this notebook.

[twiecki]

Closed via #345.

[aloctavodia]

Totally agree about focusing on hand coded model here even more we could have the handcoded version here and the formula version in bambi doc and mutually crossrefrence notebooks.