Samuel Hoffman

Hi @adrinjalali. Apologies for poor communication about this. I’m reopening this issue because, after feedback from many other users, it sounds like stronger scikit-learn support (and probably pandas to a...

We had a good conversation about this today. Here’s a summary of what we discussed and next steps: Challenges/workarounds: - @adrinjalali mentioned some relevant feature additions under review on scikit-learn:...

I finally started some work on this in the `sklearn-compat` branch. In the [README](https://github.com/IBM/AIF360/tree/sklearn-compat/aif360/sklearn) for `aif360.sklearn` I put together a rough to-do list/roadmap. A lot of it is straightforward and...

You're correct that there is no inverse_transform function available. However, the original data (`train` or `test` in the demo) is unchanged by the `fit_transform()` function so you can still use...

@vsoma11, thanks for your work! Can we make the consistency_score changes a separate PR, though, since it tracks a different issue? Also, see my other comments above.

See also: [demo_reject_option_classification.ipynb](https://github.com/Trusted-AI/AIF360/blob/master/examples/sklearn/demo_reject_option_classification.ipynb) I'm thinking it should look vaguely like [DecisionBoundaryDisplay](https://scikit-learn.org/stable/modules/generated/sklearn.inspection.DecisionBoundaryDisplay.html) where there is a `from_estimator` classmethod which takes a GridSearchCV class with 2 scorers and looks at the `cv_results_`

Can you elaborate on the shortcomings of the current method?

Is this what you're looking for (see the "Load a custom dataset" section)? https://github.com/Trusted-AI/AIF360/blob/master/examples/sklearn/monthly_bee_datasets_metrics.ipynb

See also: [demo_new_features.ipynb](https://github.com/Trusted-AI/AIF360/blob/master/examples/sklearn/demo_new_features.ipynb) and [On Fairness and Calibration](https://arxiv.org/pdf/1709.02012.pdf) (and possibly #365) Notebooks should show off a use-case for the algorithm on a known dataset, explain other ways to use it/optional...

A couple thoughts: - Can these just be functions? Is there a use case that requires it to be a class? This would make it more in line with `fetch_openml`...