Skip to content

Evaluation & Quality

Classification threshold

The score at which an AI changes its decision, such as choosing to flag a message.

Example

A fraud system flags transactions when their risk score exceeds a chosen value.

Why people use it

It lets teams choose when a score should lead to action.

What you'll hear

“At what score should we flag the transaction?”

What this means for you

Choose the cutoff using the workflow's tradeoff between missed cases and false alarms.

Can you control it?

Developer-only

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

Is the default decision cutoff always appropriate?
No. Different error costs can justify different decision cutoffs.
What happens when we lower the cutoff?
More cases will usually be flagged, including both useful detections and additional false alarms.
Can the cutoff differ between tasks?
Yes. Missing a serious problem may matter more in one task than raising an unnecessary alarm.

Related terms

Still have questions?

Up to 500 characters.

Ask LATHIC about AI. Relevant glossary entries may be included.

Your question, the glossary entries it matches, and a rotating pseudonymous identifier go to Microsoft Azure’s OpenAI service through Vercel AI Gateway to generate an answer. Zero retention and no training are required of the provider, and LATHIC does not save your question or answer. Privacy Notice