accuracy

Terms from Artificial Intelligence: humans at the heart of algorithms

The glossary is being gradually proof checked, but may have typos and misspellings.

Accuracy is the difference between the output or prediction of an algorithm and the true value. For numerical results it is often measured using the mean square error (the average of the square of the difference) or absolute difference. For classifications there are several different kinds of accuracy measure that are important, including precision and recall for binary decisions, so the word 'accuracy' can be ambiguous.

Used in glossary entries: accuracy measure, mean square error, precision, recall