How to use the Confusion Matrix Calculator
- 1. Enter the four countsType true positives, false positives, false negatives and true negatives.
- 2. CalculatePress the button to get every metric with its formula.
- 3. Pick the right metricUse the notes to choose the metric that fits your problem.
Formulas and a worked example
- Precision = TP ÷ (TP + FP): of everything flagged positive, how much was right.
- Recall = TP ÷ (TP + FN): of all real positives, how many were caught.
- F1 = 2 × precision × recall ÷ (precision + recall), the harmonic mean of the two.
- MCC = (TP·TN − FP·FN) ÷ √((TP+FP)(TP+FN)(TN+FP)(TN+FN)).
Example: a spam filter on 100 emails gets TP = 50, FP = 10, FN = 5 and TN = 35. Accuracy is 85%. Precision is 50 ÷ 60 = 0.8333 and recall is 50 ÷ 55 = 0.9091, so F1 = 0.8696. Specificity is 35 ÷ 45 = 0.7778 and MCC is 0.6975.
Which metric should you report?
Accuracy is fine when classes are balanced and both mistakes cost the same. Real problems rarely look like that. In fraud or disease detection, positives are rare, so a model that always says "no" can score 99% accuracy and still be useless.
- Recall first when missing a positive is costly: cancer screening, fraud, safety defects.
- Precision first when false alarms are costly: spam filters that might hide real mail, or legal flags.
- F1 when you need one number that balances both.
- MCC uses all four cells and stays honest on imbalanced data; many data scientists prefer it.
These match scikit-learn's classification_report for the positive class. To see how the numbers move as you change the decision threshold, try the ROC curve and AUC calculator.
Frequently asked questions
What does undefined mean for a metric?
The formula would divide by zero. For example, precision is undefined if the model never predicts positive.
Is the positive class always the rare one?
By convention you pick the class you care about catching as positive, which is often the rare one.
Is the Confusion Matrix Calculator free?
Yes, the Confusion Matrix Calculator is free. The first time you see your first result, you are asked for your name and WhatsApp number once. After that, this and every other Uncodemy tool opens straight away without the form.