AIExplainer
Machine Learning Intermediate 2 min read

What is mean average precision at k?

A metric used to evaluate the performance of information retrieval and ranking systems

Stands for: MAP@k

Mean Average Precision at k is a measure that calculates the average precision of a system's top k recommendations, where precision is the number of relevant items retrieved divided by the total number of items retrieved

Imagine you're searching for a specific type of book on a website, and the website shows you its top 5 recommendations. Mean Average Precision at k would measure how accurate those top 5 recommendations are, by checking how many of them are actually the type of book you're looking for

A music streaming service uses MAP@k to evaluate the performance of its 'recommended songs' feature, by checking how many of the top 10 recommended songs are actually liked by the user

MAP@k is used to evaluate the performance of systems that rank items, such as search engines, recommendation systems, and machine learning models

Some people think that MAP@k is the same as precision, but it's actually a more nuanced metric that takes into account the ranking of the items

Mean Average Precision at k originated in the field of information retrieval, where it was used to evaluate the performance of search engines

MAP Mean Average Precision

Three products for different needs — explore what’s relevant to you.