Question 5
Domain 6You are working on a binary classification ML algorithm that detects whether an image of a classified scanned document contains a company’s logo. In the dataset, 96% of examples don’t have the logo, so the dataset is very skewed. Which metric would give you the most confidence in your model?
Correct answer: D
Explanation
F1 score is best for a skewed binary classification dataset because it balances precision and recall instead of being dominated by the majority class. The guide emphasizes “metrics interpretation,” and F1 is defined as the harmonic mean of precision and recall, making it more informative than accuracy when “96% of examples don’t have the logo.”
Why each option is right or wrong
A. Precision
B. Recall
C. Accuracy
D. F1 score
In a binary classification problem with a 96% negative class, plain accuracy can be misleading because a trivial classifier that always predicts “no logo” would already score 96%. The F1 score is the harmonic mean of precision and recall, so it directly penalizes both false positives and false negatives; that makes it the right metric when the positive class is rare and the exam is testing metrics interpretation rather than majority-class performance.