F1 score

F1 Score

The harmonic mean of precision and recall — a single metric that balances both, where 1.0 is perfect and 0.0 is worst.

When you'd see it: Binary classification model evaluation, especially when classes are imbalanced and both precision and recall matter. Common in NLP tasks, medical classification, and fraud detection.

Why it matters: F1 collapses precision and recall into one number, which is useful when comparing models or tuning a threshold. The harmonic mean penalizes extreme imbalance — a model with perfect precision but zero recall has F1 of 0, not 0.5.

Common mistakes: Using F1 when precision and recall have very different business costs. F1 weights them equally — if catching misses (recall) matters far more than avoiding false alarms (precision), optimize directly for recall at an acceptable precision level, or use a weighted F1.

Related precision, recall, confusion matrix, AUC