easy5 pts · first answer countsmodel-monitoringprediction-drift

Watching prediction distribution

Labels for a churn model arrive after 30 days. What can you monitor today that often signals trouble early?

A.The training loss curve from last yearB.The number of model files in storageC.The CPU type of the training machineD.The distribution of predicted scores

Discussion 0

Join free to post in the discussion.

Watching prediction distribution

Discussion 0

Related questions