Anyone have any thoughts on when it7;s best to use se ML v. Scorecards?

This blog compares predicted probabilities vs. observed proportions at the feature/predictor level. The example finds that the is consistently under-estimating good credit risk across all bins of this predictor while the risk demonstrates less discrepancy between the estimated and observed outcome.

Source link
thanks you RSS link


Please enter your comment!
Please enter your name here