search

TuringLang / TuringGLM.jl

Bayesian Generalized Linear models using `@formula` syntax.
https://turinglang.org/TuringGLM.jl/dev
MIT License
70 stars 7 forks source link

Slightly Broader Name? #16

Closed ParadaCarleton closed 2 years ago

ParadaCarleton commented 2 years ago

Might be worth it to give a broader name than TuringGLM in case we want to expand a bit past GLMs. For instance, being able to specify a generalized additive model where some components are estimated with a Gaussian process regression would be a very neat feature.

phipsgabler commented 2 years ago

I have to say I have wondered about this as well. The Stan equivalent is rstanarm, "advanced regression modelling" if I'm not mistaken -- TuringAdvancedRegressionModelling?

storopoli commented 2 years ago

What acronym or names do you guys suggest? TuringAdvancedRegressionModelling is huge.

phipsgabler commented 2 years ago

JuTurArm? :laughing:

Seriously, I don't really have any perfect options myself. TuringARM would likely run afoul of the package naming guidelines. Maybe just TuringRegressionModels? Anyway, I don't want to strongly push anything here, it's really your choice.

storopoli commented 2 years ago

TuringRegressionModels is a good name but also huge.

cpfiffer commented 2 years ago

It should probably be short -- TuringGLM gets the point across moderately well, but I agree with Carlos that it falls short of the potential in here. TuringRegression is a good candidate IMO, could also suggest TuringClustering or other common modelling framework names.

ParadaCarleton commented 2 years ago

Simple Modeling or Regression with Turing: SMORT

(Yes, I did fill this in as a backronym after thinking it would be funny to have a package named SMORT)

cpfiffer commented 2 years ago

Simple Modeling or Regression with Turing: SMORT

(Yes, I did fill this in as a backronym after thinking it would be funny to have a package named SMORT)

Lol I appreciate the vigor but it's a downvote from me.

ParadaCarleton commented 2 years ago

Understandable, have a nice day

storopoli commented 2 years ago

We should either have a reference from a Julia ecosystem package (this is why I used the *GLM.jl) or from Bayesian modeling in other languages (brms or rstanarm, but please no bambi).

rikhuijzer commented 2 years ago

TuringGLM is being registered currently.