Testing and improving Agent recognition: selection of Confidence threshold

You can select the most optimal

@Confidence Threshold⁠

value for the

@NLU⁠

@Slot⁠

as follows:

If relevant

@Request⁠

s are recognized correctly and fall into the necessary

@Intent⁠

s, and irrelevant

@Request⁠

s end up not in

@fallback⁠

, but in

@Intent⁠

Increase the

@Confidence Threshold⁠

setting of the

@NLU⁠

@Slot⁠

Retrain the

@Agent⁠

and test again with the same list of questions.

If relevant

@Request⁠

s are not recognized and end up in

@fallback⁠

, just like irrelevant

@Request⁠

Lower the NLU Slot

@Confidence Threshold⁠

setting.

Retrain the

@Agent⁠

and test again with the same list of questions.

If after these steps the result is still unsatisfactory, you need to work with the

@Training Dataset⁠

and the list of

@Intent⁠

Natural language recognition⁠

If the process loops and the optimal

@NLU⁠

@Slot⁠

@Confidence Threshold⁠

value is found, the process terminates.

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.