What is the measure of how many positive instances a model identifies compared to all relevant instances called?

The measure that evaluates how many positive instances a model identifies compared to all relevant instances is known as sensitivity, often referred to as the true positive rate. It specifically assesses the model's ability to correctly identify those instances that are truly positive within a given dataset. In other words, sensitivity answers the question: of all actual positive cases, how many did the model correctly predict as positive?

This concept is particularly important in scenarios where the identification of positive cases is critical, such as in medical diagnoses, fraud detection, or any context where missing a positive instance could lead to serious consequences.

Precision, in contrast, focuses on the ratio of true positive instances among the instances that the model predicted as positive, thereby addressing a different aspect of model performance. Specificity measures the proportion of true negatives among all actual negatives, while accuracy provides an overall performance metric that encompasses both true positive and true negative predictions compared to the total number of instances.

Understanding sensitivity is vital for evaluating a model's effectiveness when the goal is to minimize false negatives and ensure that most actual positive instances are identified.