Confusion Matrix Calculator

What this tool does

Calculate confusion matrix using your data. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

How to use the Confusion Matrix Calculator

1. Enter or select true positives (tp).
2. Enter or select false positives (fp).
3. Enter or select false negatives (fn).
4. Enter or select true negatives (tn).
5. Read the calculated result; change any measurement to compare alternatives.

Formula

accuracy = (TP+TN)/total; precision = TP/(TP+FP); recall/sensitivity = TP/(TP+FN); specificity = TN/(TN+FP); F1 = 2·precision·recall/(precision+recall)

tp

True positives (TP)

fp

False positives (FP)

fn

False negatives (FN)

tn

True negatives (TN)

Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

Worked example

For confusion matrix calculator, the following measurements illustrate the exact method: True positives (TP): 80; False positives (FP): 10; False negatives (FN): 20; True negatives (TN): 90.

Results explained

Accuracy (%)

Accuracy (%) from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

Precision (%)

Precision (%) from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

Recall / sensitivity (%)

Recall / sensitivity (%) from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

Specificity (%)

Specificity (%) from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

F1 score

F1 score from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

Negative predictive value (%)

Negative predictive value (%) from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

False positive rate (%)

False positive rate (%) from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

Balanced accuracy (%)

Balanced accuracy (%) from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

Total cases

Total cases from the formula above. Accuracy alone misleads on imbalanced data — a model that never flags a 1%-prevalence condition is 99% accurate and useless. Read precision, recall and specificity together.

Frequently asked questions