Recall

noun

say it: “REE-kawl” /ˈriˌkɔl/

Of all the items that really are positive, the share a model correctly found; also called sensitivity or the true positive rate.

What is recall for a classifier?

Recall is, of the examples that really are the positive class, the share the model caught: TP / (TP + FN). Low recall means misses slip through. It's also called sensitivity and the true positive rate, the name fairness checks use when they compare it group by group.

from sklearn.metrics import recall_score
actual = [1, 0, 0, 1, 1, 0]    # 1 = fraud
guessed = [1, 0, 1, 0, 1, 0]
print(round(recall_score(actual, guessed), 3))   # 0.667: 2 of 3 frauds caught

In a sentence

  • “Lowering the threshold raised recall and brought more false alarms.”
  • “A recall of 0.4 meant the check missed six frauds in ten.”

Level: Journeyman · More AI Basics terms at this level

Updated