Refer to the exhibit: Based upon the comparative ROC plot for two competing models, which is the champion model and why?

Question

Options

• ACandidate 1, because the area outside the curve is greater
• BCandidate 2, because the area under the curve is greater
• CCandidate 1, because it is closer to the diagonal reference curve
• DCandidate 2, because it shows less over fit than Candidate 1

• A
  8% (4)
• B
  76% (39)
• C
  14% (7)
• D
  2% (1)

Explanation

Option B is correct because in ROC (Receiver Operating Characteristic) analysis, a higher Area Under the Curve (AUC) indicates better model discrimination - the ability to distinguish between positive and negative classes. Candidate 2's curve bows further toward the top-left corner, meaning it achieves higher true positive rates at lower false positive rates, making it the superior champion model.

Why the distractors are wrong:

• A is backwards - the area outside the curve is irrelevant; more area under the curve is what matters.
• C is the opposite of correct - a model closer to the diagonal reference line performs no better than random chance (AUC ≈ 0.5), which is the worst possible outcome.
• D introduces overfitting as a criterion, which cannot be determined from a single ROC plot alone; ROC curves don't directly visualize overfitting.

Memory tip: Think of the ROC curve as a "mountain" - the model with the higher mountain (more area underneath) wins. The diagonal is flat ground (random guessing), so you always want to be as far above the diagonal as possible.

No community discussion yet for this question.