Confusion Matrix Calculator

Compute precision, recall, F1, specificity, MCC and more from a confusion matrix or pasted predictions

Binary confusion matrix

Predicted → Positive | Negative

True Positive

actual +, predicted +

False Negative

actual +, predicted −

False Positive

actual −, predicted +

True Negative

actual −, predicted −

Total samples: 200

Metrics

Accuracy 87.50%

Precision 89.47%

Recall (sensitivity) 85.00%

F1 score 0.8718

Specificity 90.00%

False positive rate 10.00%

NPV 85.71%

Balanced accuracy 87.50%

MCC 0.7509

Evaluating classifiers properly

Accuracy alone hides failure on imbalanced datasets — a model predicting "negative" for everyone scores 99% accuracy on 1%-positive data. That's why precision, recall, F1, and MCC matter. Enter your four confusion-matrix counts, or paste raw actual/predicted pairs from your model's output and the tool builds the matrix for you. Hover any metric for its formula. Useful for ML coursework, Kaggle, and model reports.

Your data stays private. All processing happens locally in your browser. No data is collected, stored, or sent to any server.