KenSciResearch/fairMLHealth

AB#11757 - Bug: model_comparison.compare_measures should break with multiple protected attributes (but currently does not)

Closed this issue · 0 comments

Because certain measures from AIF360 and FairLearn can only be generated for a single protected attribute at a time, the function rejects attempts to compare measures for multiple protected attributes. Ideally, all measures should be available for multiple protected attributes, but this will take time to develop.