ROC-AUC
also: AUC · area under the curve
A single number (0.5–1.0) summarizing how well a classifier ranks positives above negatives across all thresholds.
ROC-AUC is the area under the ROC curve, which plots true-positive rate against false-positive rate as you sweep the classification threshold. The AUC summarizes, in one number from 0.5 (random) to 1.0 (perfect), how well the model ranks a random positive above a random negative — independent of any specific threshold.
Worked example: an AUC of 0.85 means that, given one positive and one negative example at random, the model scores the positive higher 85% of the time — useful for comparing models before you pick an operating threshold. Gotcha: ROC-AUC can look deceptively good on heavily imbalanced data (the false-positive rate has a huge negative denominator), so for rare-positive problems the precision-recall AUC is more informative; and a high AUC says the ranking is good, not that any single threshold’s precision or recall is acceptable.