Explainability Report
Model explained: random_forest
Global SHAP Feature Importance
| feature | mean_abs_shap |
|---|---|
| numeric__MonthlyIncome | 0.0436395 |
| numeric__Age | 0.0361074 |
| categorical__OverTime_No | 0.0349845 |
| numeric__StockOptionLevel | 0.0341481 |
| categorical__OverTime_Yes | 0.0323281 |
| numeric__YearsAtCompany | 0.0289546 |
| numeric__YearsWithCurrManager | 0.026743 |
| numeric__TotalWorkingYears | 0.024671 |
| numeric__NumCompaniesWorked | 0.0221327 |
| numeric__EnvironmentSatisfaction | 0.0189485 |
| numeric__DistanceFromHome | 0.0176427 |
| numeric__DailyRate | 0.0169811 |
| numeric__HourlyRate | 0.0167365 |
| numeric__JobLevel | 0.0148483 |
| numeric__MonthlyRate | 0.013781 |
| categorical__MaritalStatus_Single | 0.0134189 |
| numeric__YearsInCurrentRole | 0.0132743 |
| numeric__JobSatisfaction | 0.0127182 |
| numeric__RelationshipSatisfaction | 0.0121162 |
| numeric__PercentSalaryHike | 0.0115106 |
| numeric__WorkLifeBalance | 0.0106025 |
| categorical__Department_Sales | 0.0100086 |
| numeric__YearsSinceLastPromotion | 0.00986592 |
| categorical__JobRole_Sales Executive | 0.00930991 |
| numeric__JobInvolvement | 0.00902062 |
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.