In model evaluation, precision is defined as what?

In model evaluation, precision is defined as what?

• A. The proportion of actual positives that are predicted correctly.
• B. The proportion of predicted positives that are correct.
• C. The proportion of predicted negatives that are correct.
• D. The proportion of all predictions that are positive.

Answer: (B)

Precision measures how reliable the model’s positive predictions are. It answers: of all instances the model labeled as positive, what fraction truly belongs to the positive class? It’s calculated as true positives divided by the sum of true positives and false positives. For example, if eight items are predicted positive and six are actually positive while two are not, the precision is 6/8 = 0.75. This focuses on the trustworthiness of positive predictions, not on finding all positives. The other ideas describe different notions: one looks at how many actual positives were captured (recall), another at how many predicted negatives are correct (negative predictive value), and another at the overall rate of positive predictions, which doesn’t measure accuracy.

Precision measures how reliable the model’s positive predictions are. It answers: of all instances the model labeled as positive, what fraction truly belongs to the positive class? It’s calculated as true positives divided by the sum of true positives and false positives. For example, if eight items are predicted positive and six are actually positive while two are not, the precision is 6/8 = 0.75. This focuses on the trustworthiness of positive predictions, not on finding all positives. The other ideas describe different notions: one looks at how many actual positives were captured (recall), another at how many predicted negatives are correct (negative predictive value), and another at the overall rate of positive predictions, which doesn’t measure accuracy.