search

koaning / scikit-lego

Extra blocks for scikit-learn pipelines.
https://koaning.github.io/scikit-lego/
MIT License
1.26k stars 116 forks source link

`decomposition.umap_reconstruction.UMAPOutlierDetection` #653

Closed FBruzzesi closed 2 months ago

FBruzzesi commented 5 months ago

@anopsy created separate issues to link directly PRs

anopsy commented 5 months ago

Thanks! I'll also repeat my question:

what would be a more elegant example - applying them on ndrrays or a dataframe?

I noticed that some examples in sklego are based on dataframes and some on numpy arrays or ndarrays. I noticed that scikit-learn mostly uses arrays in their examples. Let me know what do you think, maybe I'm overthinking that

FBruzzesi commented 5 months ago

I would suggest to use arrays as well wherever possible. Dataframes may be more interpretable but they are more in the direction of skrub philosophy

anopsy commented 5 months ago

Coolio! That's more of an idea for the future, but when all examples are added, maybe it would be good to base them all on arrays then. Just sayin'

FBruzzesi commented 2 months ago

Closed by #654