Society of Actuaries (SOA) PA Practice Exam

What does precision measure in a confusion matrix?

1. True negatives among all negative cases

2. True positives among all predicted positives

3. True positives among all negatives

4. Negative rate among all predictions

The correct answer is: True positives among all predicted positives

Precision measures the proportion of true positive predictions made by a model out of all the positive predictions it has made. In the context of a confusion matrix, this is defined as the ratio of true positives to the sum of true positives and false positives. Therefore, when we say precision, we are referring specifically to the accuracy of the positive predictions. By focusing on the positive predictions, precision helps gauge how many of those identified as positive are actually correct, which is particularly important in cases where the cost of false positives is high. This metric is crucial in evaluating models in fields such as healthcare or fraud detection, where it is vital not to incorrectly classify a negative case as positive. The other options provide insights into different metrics. For instance, true negatives among all negative cases relates to specificity, while true positives among all negatives pertains to recall, and the negative rate among all predictions refers to the overall rate of negative predictions made by the model. These metrics, while valuable, indicate different aspects of a model's performance and do not directly measure precision.