A data scientist must choose between a model with AUC=0.82 and another with AUC=0.79. However, the business requires maximizing recall at a minimum precision of 0.70. Which approach best addresses this requirement?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — b is correct because for imbalanced problems with specific precision requirements, the Precision-Recall curve (not the ROC-AUC) is the right evaluation tool. The correct approach is to read off each model's recall at the precision=0.70 point from their respective PR curves and select the model with higher recall at that operating point.
Full explanation below image
Full Explanation
B is correct because for imbalanced problems with specific precision requirements, the Precision-Recall curve (not the ROC-AUC) is the right evaluation tool. The correct approach is to read off each model's recall at the precision=0.70 point from their respective PR curves and select the model with higher recall at that operating point. A is wrong because AUC summarizes across all thresholds and a higher AUC does not guarantee better performance at a specific precision constraint. C is wrong — lower AUC does not imply better calibration. D is wrong because threshold=0.70 applies to predicted probability, not precision.