📏 Precision–Recall AUC (PR-AUC)

Precision–Recall AUC (PR-AUC)#

Area under the precision-recall curve, informative under class imbalance.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What it is#

PR-AUC is the area under the precision–recall curve, which plots precision against recall across thresholds — also called Average Precision (AP), the mean precision over all recall levels. It ranges \([0, 1]\), higher is better.

Why it’s imbalance-friendly#

Unlike ROC-AUC, PR-AUC ignores true negatives entirely and focuses on the positive class, so it stays informative when positives are rare. Its baseline also shifts with prevalence — random guessing scores the positive-class ratio (0.5 when balanced, 0.01 at 1% positive), not a fixed 0.5.

When to use it#

Reach for PR-AUC on highly imbalanced problems where finding the minority positive is the goal — fraud, rare-disease, anomaly detection — where a high ROC-AUC can be misleading. Report it alongside ROC-AUC for the full picture.

Theme: Classification & Averaging Metrics · All terminology

Hint

Mind map — connected ideas

ROC-AUC (Receiver Operating Characteristic – Area Under Curve, = AUROC) · Precision (a.k.a. Positive Predictive Value, PPV) · Binary Classification · Multiclass AUROC · Classification Probability · Macro AUC

Hint

More in Classification & Averaging Metrics

Accuracy · AUC (Area Under the Curve) · Average Precision (AP) · Binary Classification · Classification Probability · Discriminatory Power · F1-score · Gini Coefficient · Harmonic Mean · Log Loss (also called Logarithmic Loss or Cross-Entropy Loss) · Macro AUC · Macro AUROC (Macro-Averaged AUROC) · Macro Averaging · Macro F1