Artificial intelligence · University
Retrieval recall on a labeled evidence set
The reference relevance labels and query set define the evaluation; recall does not measure answer correctness.
Method and conditions
The reference relevance labels and query set define the evaluation; recall does not measure answer correctness.
recall = relevant items retrieved / all relevant items
Worked examples
A labeled query has 13 relevant documents. A retriever finds 9 of them. Calculate retrieval recall. Round the final result to 4 decimal places if needed.
Use 9/13. Duplicated retrieved documents are not counted twice.
0.6923
A labeled query has 6 relevant documents. A retriever finds 2 of them. Calculate retrieval recall. Round the final result to 4 decimal places if needed.
Use 2/6. Duplicated retrieved documents are not counted twice.
0.3333
Try it
A labeled query has 36 relevant documents. A retriever finds 21 of them. Calculate retrieval recall. Round the final result to 4 decimal places if needed.
A labeled query has 12 relevant documents. A retriever finds 5 of them. Calculate retrieval recall. Round the final result to 4 decimal places if needed.
Check your work
Answers and marking notes
- 0.5833
Use 21/36. Duplicated retrieved documents are not counted twice.
- 0.4167
Use 5/12. Duplicated retrieved documents are not counted twice.
Further reading and review references
These are reference links, not copied exercises, institutional endorsement or proof of independent review.
AI-assisted preview; independent educational review required. Level labels describe prerequisites, not accreditation or official curriculum alignment.