During model monitoring in SAS, a data scientist observes that the Gini coefficient for a credit risk model has declined from 0.62 at deployment to 0.41 six months later. What does this indicate?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — when the Gini coefficient drops from 0.62 to 0.41, your model's discriminatory power has fallen sharply. That means performance degradation, likely from concept drift or data drift.
Full Explanation
The correct indication is performance degradation, likely from concept drift or data drift. The Gini coefficient, twice the area under the receiver operating characteristic curve minus one, measures discriminatory power. The drop from 0.62 to 0.41 is significant, showing reduced ability to separate good from bad borrowers. Concept drift means predictor-outcome relationships changed; data drift means input distributions changed. The idea that lower Gini coefficients indicate better calibration is wrong: lower Gini means weaker discrimination. The idea that this decline is expected and acceptable is wrong: the 21-point decline is substantial and warrants investigation. The idea that retraining with fewer variables would reduce the Gini coefficient further is wrong: reducing variables would not address the root cause of drift.