ROC Curve

The ROC curve plots TPR against FPR, showing a classifier's tradeoff across thresholds, with AUC quantifying overall performance.

Purpose

To evaluate a binary classifier's overall performance across thresholds.

Mechanism

TPR=TP/(TP+FN) (recall), FPR=FP/(FP+TN). The diagonal is a random classifier. AUC is the area under the ROC curve: 0.5 is random, 1.0 perfect. AUC is robust to class imbalance.

Use cases

Classifier evaluation
Threshold selection
Imbalanced data

Related terms

AUC

AUC is the area under the ROC curve, equivalent to the probability that a random positive sample scores higher than a negative one, measuring a binary classifier's overall discrimination.

GLOSSARY

References

Wikipedia — ROC curve
Fawcett 2006 — ROC Introduction