Evaluation & Quality
Recall
Of all truly positive items, the fraction the model successfully found.
Example
Of all actual phishing emails, recall measures how many the filter found.
Why it matters
It helps teams see whether a detector is finding enough of the problems that exist.
You'll hear it
“How many real problems escaped the check?”
Can you control it?
Set by the team
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Can recall improve while false alarms increase?
- Yes. Flagging more cases may catch additional problems while also catching harmless ones.
- Can recall be calculated without knowing the real cases?
- Not reliably. The comparison needs a trustworthy account of what should have been found.
- Does perfect recall guarantee a useful detector?
- No. A system that flags everything may miss nothing but create an unusable number of false alarms.