Recall is the number of the correct predictions of the positive class divided by the number of predictions that should have been classified in the positive class. TP/(TP + FN)
Card 1 of 259. Answer: Recall is the number of the correct predictions of the positive class divided by the number of predictions that should have been classified in the positive class. TP/(TP + FN)
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
Recall, also called sensitivity or the true positive rate, is the share of actual positives the model finds: recall = TP / (TP + FN). It answers "of everything that was really positive, how much did we catch?"
Example: there are 50 fraudulent transactions and the model flags 40 of them. Recall is 40/50 = 0.8, no matter how many legitimate transactions it also flagged by mistake. That is why recall is always reported with precision, which counts those false alarms.
Prioritize recall when missing a positive is expensive, as in cancer screening or fraud detection. You raise it by lowering the decision threshold, which usually lowers precision. Predicting everything as positive gives a recall of 1.0 and is useless, which is the tradeoff the F1 score and the precision-recall curve capture.
