Once a model looks good enough, a slice that never guided a nudge is used to check the agreed ML functional-performance criteria and is compared with the validation numbers. If the held-out score is much worse, the team returns to model generation or even to data prep. Which workflow step is that?

Short Explanation

A slice that never guided a nudge, used to check the agreed ML functional-performance criteria and compared with the validation numbers, is test-the-model. If the held-out score is much worse, go back to model generation or even data prep. That slice is not the validation set.

Full Explanation

The independent test-dataset step checks agreed criteria and the go-back rule when test is much worse than evaluation. Timing-to-predict may be checked here too. Validation, Chapter 6, and deploy re-pack are different steps.

×