Calculator Input
Enter class names, true positives, and false negatives. The result stays above this form after submission.
Formula Used
Recall for one class:
Recall_i = TP_i / (TP_i + FN_i)
Macro recall:
Macro Recall = (Recall_1 + Recall_2 + ... + Recall_n) / n
Micro recall:
Micro Recall = Sum(TP) / Sum(TP + FN)
Macro recall gives every class equal importance. It helps when class balance matters and weak minority classes should not disappear inside overall accuracy.
How to Use This Calculator
- Add one card for each target class.
- Enter a class label for easier reading.
- Fill in true positives for that class.
- Fill in false negatives for that class.
- Click Calculate Macro Recall.
- Review macro recall, micro recall, spread, and the chart.
- Download the summary as CSV or PDF if needed.
Example Data Table
This sample shows how the calculator handles four classes.
| Class | True Positives | False Negatives | Recall |
|---|---|---|---|
| Positive | 86 | 14 | 86.00% |
| Neutral | 72 | 18 | 80.00% |
| Negative | 91 | 9 | 91.00% |
| Mixed | 64 | 16 | 80.00% |
Frequently Asked Questions
1. What does macro recall measure?
Macro recall averages recall across classes equally. It shows how well a model finds each class, without letting large classes dominate the score.
2. Why use macro recall instead of accuracy?
Accuracy can look strong even when smaller classes perform poorly. Macro recall reveals whether every class gets fair detection, which matters in imbalanced datasets.
3. What is the difference between macro and micro recall?
Macro recall gives each class equal weight. Micro recall combines all true positives and false negatives first, so larger classes influence the result more strongly.
4. What happens when a class has zero support?
If TP plus FN equals zero, recall is undefined for that class. This calculator excludes that class from macro recall and shows a warning note.
5. Can I use decimals for TP and FN?
Yes. The calculator accepts decimal values. That can help when inputs come from averaged folds, normalized counts, or analytical simulations.
6. Why is class spread useful?
Recall spread shows the difference between the strongest and weakest class. A smaller spread usually means more balanced model sensitivity across categories.
7. When should I monitor macro recall closely?
Monitor macro recall when false negatives matter for every class. It is especially useful in medical screening, fraud detection, moderation, and rare-event classification.
8. Does macro recall work for multiclass problems?
Yes. Macro recall is commonly used for multiclass evaluation. It summarizes per-class sensitivity in one value while keeping class balance visible.