Which performance measure quantifies the proportion of actual positives correctly identified by a model?

The measure that quantifies the proportion of actual positives correctly identified by a model is sensitivity, also known as recall. Sensitivity is defined as the ratio of true positives to the sum of true positives and false negatives. This means it specifically focuses on how well the model can identify positive instances among all the actual positive instances present in the dataset.

In the context of performance evaluation for classification models, sensitivity is crucial for applications where identifying positive cases is particularly important, such as in medical diagnoses, where failing to identify a disease can have significant consequences.

The other measures in the options each serve different purposes:

• Accuracy assesses the overall correctness of the model across both positive and negative cases but does not provide insight into how well positives are identified versus negatives.

• Specificity measures the proportion of true negatives correctly identified by the model, which focuses on the negative instances rather than the positives.

• Precision evaluates the proportion of true positives out of all instances predicted as positive, but it does not consider false negatives.

Overall, sensitivity is the most relevant measure for understanding the effectiveness of a model in identifying actual positive cases.