search

NannyML / The-Little-Book-of-ML-Metrics

The book every data scientist needs on their desk.
https://www.nannyml.com/metrics
777 stars 60 forks source link

Equal Opportunity Difference #115

Closed pratikwatwani closed 3 weeks ago

pratikwatwani commented 1 month ago

Metric's name Equal Opportunity Difference

Metric's category Bias & Fairness/Business

Metrics formula Equal Opportunity Difference= TPR(group 1) −TPR(group 2)

where TPR (True Positive Rate) = True Positives/(True Positives + False Positives) ​group 1 and 2 indicate different demographic groups

Describe the metrics use cases, and any relevant references. EOD is used to assess if an ML model’s performance is consistent across different demographic groups (male vs female, low vs high income level, race etc) and ensuring that the model offers equal access to positive outcomes.

santiviquez commented 3 weeks ago

I believe this is duplicated https://github.com/NannyML/The-Little-Book-of-ML-Metrics/issues/84

pratikwatwani commented 3 weeks ago

Yes, you are right, I'll close it.