Explainability Report
Model explained: logistic_regression
Global SHAP Feature Importance
| feature | mean_abs_shap |
|---|---|
| numeric__JobLevel | 0.549735 |
| numeric__MonthlyIncome | 0.471733 |
| numeric__TotalWorkingYears | 0.424355 |
| numeric__YearsSinceLastPromotion | 0.423696 |
| numeric__NumCompaniesWorked | 0.405947 |
| categorical__OverTime_No | 0.359265 |
| numeric__JobSatisfaction | 0.346377 |
| numeric__EnvironmentSatisfaction | 0.337649 |
| categorical__OverTime_Yes | 0.306879 |
| numeric__RelationshipSatisfaction | 0.293467 |
| numeric__YearsWithCurrManager | 0.283127 |
| numeric__YearsInCurrentRole | 0.277629 |
| categorical__BusinessTravel_Travel_Frequently | 0.253191 |
| categorical__JobRole_Laboratory Technician | 0.249432 |
| numeric__DistanceFromHome | 0.231682 |
| numeric__Age | 0.231313 |
| categorical__MaritalStatus_Single | 0.230747 |
| categorical__MaritalStatus_Divorced | 0.203018 |
| numeric__JobInvolvement | 0.202091 |
| numeric__TrainingTimesLastYear | 0.196138 |
| categorical__BusinessTravel_Non-Travel | 0.175278 |
| categorical__JobRole_Sales Representative | 0.173284 |
| numeric__StockOptionLevel | 0.171638 |
| numeric__WorkLifeBalance | 0.154537 |
| numeric__PercentSalaryHike | 0.152525 |
Notes
- SHAP values explain feature contributions to model predictions.
- Global importance is computed as mean absolute SHAP value.
- Local explanations are generated for representative validation samples.