Machine Learning Large Language Models Intermediate 1 min read

What is precision at k?

A measure of the accuracy of a model's top k predictions

precision at k explained in plain English

Precision at k is a metric used to evaluate the performance of a model by calculating the proportion of relevant items among the top k items it recommends or predicts

Analogy

Imagine you're searching for a restaurant and a recommendation system gives you the top 5 suggestions. Precision at k would measure how many of those top 5 suggestions are actually good restaurants that you would like

Example

A music streaming service uses precision at k to evaluate the effectiveness of its 'top 10 songs' recommendation feature, where k is 10

How is precision at k used?

Precision at k is used in information retrieval and recommendation systems to evaluate the quality of the recommendations or predictions made by a model

Common misconceptions about precision at k

Some people confuse precision at k with recall at k, but precision at k focuses on the accuracy of the top k predictions, while recall at k focuses on the proportion of relevant items that are included in the top k predictions

History

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

Also known as

precision at top k p@k