How to use facet package with pretrained sklearn pipeline ?

[Paroag]

I'm working on outlier detection, and using sklearn pipelines in my project. I currently compute shap values for my outliers to better understand how my underlying models use my features.

I would love to dig further into explainability by exploring the synergy and redundancy as defined in facet, however I noticed in most examples that the facet package uses sklearndf as the framework for ML models.

Is there a way to use this package with standard sklearn pipeline ? Any recommandation on how to use this package is very appreciated as i'm new to it.

[j-ittner]

Hi @Paroag , not yet but we're considering enhancing the LearnerInspector to also accept native scikit-learn pipelines. If you want, create an issue & we'll keep you up to date on progress! Thanks!

[Paroag]

Thanks for the quick answer !

[j-ittner]

Update: this will be a feature in FACET 2.1, see PR #369