Recall
In plain terms
Of actual positives, the fraction the model correctly found.
Going deeper
Recall measures how many true positives a model catches, important when missing positives is costly. It answers 'of all the real yeses, how many did it find?'
Under the hood
Trades off against precision; high recall means few misses but possibly more false alarms.
Examples
Medical screening, fraud detection, search coverage, evaluation metrics.
Think of it like...
Like a disease screen judged on how few sick patients it misses.