Recall
The share of actually positive cases that the model correctly identifies as positive.
How many relevant cases are found?
Recall measures what share of actually positive cases according to the reference are correctly identified as positive. The formula is TP / (TP + FN): TP means correctly identified positives and FN missed positive cases. The positive class must be defined. If the assessment contains no actually positive cases, the denominator is zero and the ratio is undefined. Reporting must keep that situation visible.
In the fictional calculation with 50 service requests, 20 belong to technical service under the agreed assignment rule. The model identifies 18 correctly and misses 2. Recall is 18 / 20 = 90 %. For work planning, this reveals the gap among cases that should reach the technical team. It does not reveal how many other requests are also wrongly sent there; precision examines that side. Whether a setting is useful is assessed against both error types and their consequences in the specific workflow.
Sources
Related terms
- LLM evaluationLLM evaluation tests model outputs against defined tasks and criteria. One good example does not establish overall quality.
- Test data in machine learningData reserved for final assessment whose results were not used to select the model variant being assessed.
- PrecisionThe share of actually positive cases among all cases the model marks as positive.
- F1 scoreThe harmonic mean of precision and recall for the same positive class.
- Confusion matrixA table comparing reference classes with model predictions and counting correct and incorrect assignments.
- Re-rankingRe-ranking reassesses search candidates already retrieved. Relevance to the query and the limits of the initial selection both matter.
Your project with ASES
Describe your project to us. We will get back to you with an initial assessment.