In SAS Model Manager, a data scientist configures a performance monitoring job that compares model predictions to actual outcomes. Which SAS procedure is most commonly used within this monitoring workflow to generate model fit statistics?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — you want PROC ASSESS with the FITSTAT option because it's SAS's purpose-built model assessment procedure for scored data where actual outcomes are known, generating KS statistic, cumulative lift, ROC curves, and other fit statistics in Model Manager monitoring workflows. The HPLOGISTIC option with VALIDATE= dataset is for training-time validation, not post-deployment monitoring.
Full Explanation
PROC ASSESS with the FITSTAT option is correct because PROC ASSESS is SAS's purpose-built model assessment procedure that computes model performance statistics including KS statistic, cumulative lift, ROC curves, and other fit statistics on scored data where actual outcomes are known, making it the standard tool in SAS Model Manager monitoring workflows. The HPLOGISTIC option with VALIDATE= dataset is a training-time tool, not a post-deployment monitoring tool. The COMPARE option compares dataset structure and values, not model performance. The TABULATE option produces descriptive tables but not calibrated model fit statistics.