What does Recall measure in a classification model?

Recall is a key performance metric used in classification models that quantifies the ability of a model to identify all relevant instances within a dataset. Specifically, recall measures the proportion of actual positive instances that are correctly identified by the model as positive. This is crucial in contexts where missing a positive instance has significant ramifications—such as in medical diagnoses or fraud detection.

Thus, the correct answer reflects that recall captures how well the model is performing in identifying true positive cases out of all the actual positive cases present. High recall indicates that most of the positive labels are correctly predicted, making it an essential metric when the goal is to minimize false negatives.

Other choices touch on different aspects of model performance but do not accurately define recall. For example, while the total number of true positives denotes one component involved in the recall calculation, it does not provide the crucial context of proportion to total actual positives. The ratio of true positives to total predicted positives refers to precision, which is a different metric entirely. Lastly, the total number of actual negative predictions is irrelevant to measuring recall, as it strictly pertains to positive cases only.