precision

Terms from Artificial Intelligence: humans at the heart of algorithms

Page numbers are for draft copy at present; they will be replaced with correct numbers when final book is formatted. Chapter numbers are correct and will not change now.

When making a binary classification or decision, such as a disease diagnosis, precision is the probability that if we make a positive decision, it turns out out to be correct, for example, if we diagnose a person with the disease they really have it. If a decision has a high precision, then the probability of a false positive is low. Often increasing precision reduces recall; this is called the precision–recall trade-off.

Defined on page 180

Used on Chap. 9: pages 180, 181, 182, 193, 196; Chap. 18: page 447