Which metric is not typically used to evaluate classification models?

• A. Accuracy
• B. F1 Score
• C. Mean Squared Error
• D. ROC-AUC

Answer: (C)

Classification models are evaluated by how well they assign the correct category and, for some metrics, how well they rank positive cases above negatives. Mean Squared Error is a regression-style measure that computes the average squared difference between predicted numeric values and true targets. In classification, targets are categorical (often 0/1 or other labels), so the goal is to minimize misclassifications and to understand how the model performs across decision thresholds, not to minimize numeric prediction error magnitude. MSE doesn’t align with misclassification rates or threshold-based decisions, and it can give misleading signals about classifier performance. That’s why it isn’t typically used for evaluating classifiers. Instead, accuracy, F1 score, and ROC-AUC directly reflect classification performance: accuracy measures overall correctness, F1 balances precision and recall, and ROC-AUC assesses the model’s ability to discriminate between classes across all thresholds.