Accuracy

Accuracy is the most basic way to measure a classification model: the fraction of predictions it got right out of all predictions made.

Data & AI

Accuracy = correct predictions / total predictions. It's intuitive and easy to report, but it can be misleading on imbalanced data. If 99% of emails are not spam, a model that labels everything 'not spam' scores 99% accuracy while catching zero spam.

For that reason, accuracy is usually reported alongside precision, recall, and a confusion matrix to give a fuller picture of performance.

Examples

Python

from sklearn.metrics import accuracy_score

y_true = [1, 0, 1, 1]
y_pred = [1, 0, 0, 1]
print(accuracy_score(y_true, y_pred))

Output:

0.75

Learn it in a lesson

See also

Back to the Dictionary