Average Precision (AP)

Average Precision combines precision and recall into a single score by summarizing performance across different classification thresholds. It measures how effectively a model ranks correct positive predictions ahead of incorrect ones. A higher AP score generally indicates better performance. AP is particularly useful when evaluating models on datasets where positive and negative examples are unevenly distributed.

Accuracy alone can be misleading when working with imbalanced datasets. Average Precision provides a more informative view by considering both precision and recall. It helps teams compare models, evaluate detection quality, and understand how reliably a model identifies relevant instances while limiting false positives.