Skip to content
ML Academy

Introduced on day 13

Average precision (PR-AUC)

The area under the precision-recall curve, computed as a precision-weighted average over recall steps; a constant score earns the prevalence.

In detail

Average precision summarizes how pure the top of the ranking is across recall levels. Unlike ROC-AUC it moves with prevalence, because precision depends on how many negatives surround each positive, so compare it with its floor (the prevalence), not with 0.5. Different libraries compute PR-AUC by step-wise average precision or by trapezoids; state which one you use.