Neural Mastery

Evaluation

ROC AUC

ROC AUC measures ranking quality across classification thresholds using true-positive and false-positive rates.

false positive ratetrue positive ratemodelrandom
ROC AUC measures ranking quality across classification thresholds using true-positive and false-positive rates. Values shown in the worked visual are illustrative unless the labels state otherwise.

How it works

It equals the probability that a random positive receives a higher score than a random negative under standard assumptions. It can look optimistic on highly imbalanced problems.

A concrete example

Illustrative: an illustrative ROC curve has AUC 0.87 across all thresholds.

Do not confuse it with PR AUC

ROC uses false-positive rate; PR uses precision and recall and is often more revealing for rare positives.

Study the full lesson