AUC
In plain terms
A metric summarizing a classifier's performance across thresholds.
Going deeper
AUC (Area Under the ROC Curve) measures how well a model ranks positives above negatives regardless of threshold. Higher means better separation.
Under the hood
0.5 is random, 1.0 is perfect; threshold-independent and robust to class balance for ranking.
Examples
Classifier evaluation, ROC curves, ranking quality, model comparison.
Think of it like...
Like rating a sorter on how well it puts all the good apples ahead of the bad ones.