ROC Area Under the Curve (AUC)

Definition

A summary statistic derived from the ROC curve, representing the probability that a classifier will rank a randomly chosen positive instance higher than a randomly chosen negative one. AUC = 1.0 indicates perfect discrimination; AUC = 0.5 indicates no better than chance.

Why It Matters

AUC condenses the entire ROC curve into a single number that is independent of any classification threshold, making it ideal for comparing models. It is widely used in medical testing, credit scoring, and machine learning to summarise how well a model distinguishes between two classes, without committing to a specific decision boundary.

Example

A credit-scoring model achieves an AUC of 0.82, meaning there is an 82% probability that a randomly selected borrower who defaults receives a higher risk score than a randomly selected borrower who does not. This single number allows the bank to compare the model against alternatives without examining every possible threshold.

Related Terms

• Receiver Operating Characteristic (ROC) Curve
• Significance Level
• Sensitivity Testing

Software Notes

• SPSS: Analyze > Classification > ROC Curve; AUC is displayed automatically
• R: auc() from the pROC package; e.g., auc(roc(response, predictor))
• Stata: roctab displays AUC; lroc after logistic regression