Skip to content
LearnPY
scikit-learnfunction reference

accuracy_score

Measure the fraction of correctly classified rows.

Signature

accuracy_score(y_true, y_pred, normalize=True)

What it does

accuracy_score compares predicted labels with true labels. It is easy to interpret but can hide poor performance on imbalanced classes, where a majority-class baseline may already score highly.

Parameters

y_truearray_like
Ground-truth labels.
y_predarray_like
Predicted labels in the same order.
normalizebool, optionaloptional
Return a fraction when True, a count when False.

Returns

float

Example

pythonaccuracy_score example
from sklearn.metrics import accuracy_score

score = accuracy_score(y_test, predictions)
print(f"accuracy: {score:.3f}")

Watch out for

  • Accuracy alone is misleading for rare classes; also inspect precision, recall, and a confusion matrix.
  • y_true and y_pred must be aligned row by row.