Glossary term
Glossary term
Evaluation and Benchmarks
A metric for summarizing a model's performance on a single prompt that generates ranked results, such as a numbered list of book recommendations. Average precision at k is, well, the average of the precision at k values for each relevant result. The formula for average precision at k is therefore:
where:
is the number of relevant items in the list.
Contrast with recall at k.
Note: Average precision at k evaluates the output for a single prompt. Use mean average precision at k to evaluate the quality of a model's output across many different prompts.Note: Some people abbreviate average precision at k to simply average precision.
Click the icon for an example
Created for this library
A recommender team reports average precision at 10 for its homepage carousel because the top of the list is where shoppers actually engage.
A help-center search team tracks average precision at 5 on its support content because users rarely scroll past the first five results.
A music streaming team uses average precision at 20 on its daily mix to balance relevance and diversity across the listening session.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License