Concept
ROC curves and the area under the curve (AUC) as evaluation tools, and why a high AUC can still hide group-level unfairness.
One-sentence definition
A ROC curve plots true positive rate against false positive rate across every decision threshold, and AUC summarizes that curve as a single number.
What real data will illustrate it?
Plot per-group ROC curves from one of the healthcare audits (Healthcare Readmission) to show that equal overall AUC can mask different per-group curves.
Limitations / trade-offs
AUC is threshold-independent and can look strong while calibration and error rates differ sharply across groups. Good first issue for someone comfortable with matplotlib.
Hyphens only - no em dashes.
Concept
ROC curves and the area under the curve (AUC) as evaluation tools, and why a high AUC can still hide group-level unfairness.
One-sentence definition
A ROC curve plots true positive rate against false positive rate across every decision threshold, and AUC summarizes that curve as a single number.
What real data will illustrate it?
Plot per-group ROC curves from one of the healthcare audits (Healthcare Readmission) to show that equal overall AUC can mask different per-group curves.
Limitations / trade-offs
AUC is threshold-independent and can look strong while calibration and error rates differ sharply across groups. Good first issue for someone comfortable with matplotlib.
Hyphens only - no em dashes.