Calculate recall (sensitivity) from true positives and false negatives.
Recall, also called sensitivity or true positive rate (TPR), measures how many of the actual positive cases the model correctly identified: recall = TP / (TP + FN). High recall means the model misses few real positives. Recall says nothing about false positives; it's the mirror image of precision, and the two are typically reported together because improving one often comes at the expense of the other, depending on the classification threshold.
recall = TP / (TP + FN)
Prioritize recall when false negatives are costly — for example, missing a cancer diagnosis or failing to detect fraud — where catching every real positive matters more than occasionally flagging a false alarm.
Yes, recall, sensitivity, and true positive rate (TPR) are all names for the same metric: TP / (TP + FN). The term used often depends on the field — 'sensitivity' is common in medical diagnostics, 'recall' in information retrieval and ML.
Lowering the decision threshold for classifying something as positive typically increases recall (catching more true positives) but also increases false positives, reducing precision — this tradeoff is visualized by the ROC and precision-recall curves.