When computing the F1 score for an imbalanced binary classification problem, which formula is correct?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — the right choice is the one saying both the harmonic mean formula and the TP/(TP + 0.5*(FP + FN)) formula are equivalent correct formulas. You're picking the complete answer, not just one valid-looking expression.
Full Explanation
The correct choice is the one saying both the harmonic mean formula and the TP/(TP + 0.5(FP + FN)) formula are equivalent correct formulas. The harmonic mean formula, 2 (Precision * Recall) / (Precision + Recall), is standard. Substituting precision and recall into it simplifies to the TP-based formula, so both are correct. The option with only the harmonic mean formula is incomplete, and the option with only the TP-based formula is incomplete. The arithmetic mean of precision and recall is wrong because F1 uses the harmonic mean.