Precision and Recall
Definition
Precision and Recall jointly evaluate accuracy and completeness of hits.
A pair of metrics: Precision measures the proportion of truly positive cases among the positive predictions, Recall measures the proportion of found positive cases among all truly positive cases.
Simply explained
One number asks how many hits are correct; the other, how many correct cases were found.
Why it matters in practice
The term helps to select, develop, evaluate AI systems in a well-founded manner and realistically assess their technical risks.
Example from business
A document search evaluates both the relevance of the hits and how many relevant documents were found.
Limits and pitfalls
- Do not optimize a single metric in isolation or fail to report threshold and base rate.
Origin & history
Precision and Recall originate from information retrieval and classification.
Use in mid-sized companies
- Possible example: A document search evaluates both the relevance of the hits and how many relevant documents were found.
Frequently asked questions
What do Precision and Recall mean?
A pair of metrics: Precision measures the proportion of truly positive cases among the positive predictions, Recall measures the proportion of found positive cases among all truly positive cases. What mistakes should be avoided? Optimizing a single metric in isolation or failing to report threshold and base rate.
Sources
Source checked on 10.10.2026. Editorial application examples are hypothetical; no proof of real customer results.
Editorially reviewed on: