AI engineering glossary · Evaluation metrics
What is MRR (mean reciprocal rank)?
Mean reciprocal rank (also called MRR): The average of the reciprocal rank over many queries. It summarises how high, on average, the first relevant result appears. A value near 1 means the right answer is usually first. It is best when each query has one right answer.
Mean reciprocal rank: an example
Three queries have their first correct result at ranks 1, 2 and 4, so MRR = (1 + 0.5 + 0.25) / 3, about 0.58.
Related terms
More Evaluation metrics terms
- Mean squared error
- Mean squared logarithmic error
- Median absolute error
- METEOR
- Micro averaging
- Negative predictive value
- Normalised discounted cumulative gain
- Normalised mutual information
Where you learn it
Mean reciprocal rank is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.