recall

Terms from Artificial Intelligence: humans at the heart of algorithms

The glossary is being gradually proof checked, but currently has many typos and misspellings.

When making a binary classification or decision, such as a disease diagnosis, recall is the probability that if the person really has the disease, or in general if the classification really is positive, then we correctly make the decision; that is, if the person has the disease then we do diagnose it. If a decision has a high recall, then the probability of a false negative is low. Often increasing recall reduces precision; this is called the precision–recall trade-off.

Used in Chap. 9: pages 120, 121, 128, 129; Chap. 18: page 288