Algorithmic Recourse on German (10-fold cross-validation)
0.02Recourse Cost (MPS)SOGAR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SOGARClassifier Model=DNN2026.05 | 0.02 | 0 | 0.02 | 28.2 | |
| CETClassifier Model=DNN2026.05 | 0.04 | 0.44 | 0.48 | 255.3 | |
| SOGARClassifier Model=XGBoost2026.05 | 0.09 | 0.02 | 0.12 | 26.4 | |
| SOGARClassifier Model=LightGBM2026.05 | 0.11 | 0.03 | 0.14 | 41.9 | |
| CETClassifier Model=XGBoost2026.05 | 0.12 | 0.25 | 0.37 | 151.2 | |
| CETClassifier Model=LightGBM2026.05 | 0.13 | 0.31 | 0.44 | 173.8 | |
| AReSClassifier Model=DNN2026.05 | 0.17 | 0.39 | 0.56 | 1,845.3 | |
| AReSClassifier Model=LightGBM2026.05 | 0.19 | 0.65 | 0.84 | 680.4 | |
| T-CREx0.9Classifier Model=XGBoost2026.05 | 0.4 | 0.75 | 1.2 | 8.2 | |
| AReSClassifier Model=XGBoost2026.05 | 0.41 | 0.13 | 0.55 | 2,362.8 | |
| T-CREx0.9Classifier Model=LightGBM2026.05 | 0.44 | 0.72 | 1.16 | 8.25 | |
| GLANCEClassifier Model=DNN2026.05 | 0.47 | 0.01 | 0.47 | 8.5 | |
| GLANCEClassifier Model=XGBoost2026.05 | 0.5 | 0.01 | 0.5 | 8.4 | |
| GLANCEClassifier Model=LightGBM2026.05 | 0.52 | 0.01 | 0.53 | 9.3 | |
| T-CREx0.9Classifier Model=DNN2026.05 | 0.53 | 0.78 | 1.31 | 32.19 | |
| T-CREx0.99Classifier Model=LightGBM2026.05 | 0.54 | 0.61 | 1.15 | 10 | |
| T-CREx0.99Classifier Model=XGBoost2026.05 | 0.56 | 0.57 | 1.13 | 8.03 | |
| T-CREx0.99Classifier Model=DNN2026.05 | 0.6 | 0.7 | 1.25 | 19.63 | |
| GLOBE-CEClassifier Model=DNN2026.05 | 0.91 | 0.01 | 0.92 | 7.58 | |
| GLOBE-CEClassifier Model=LightGBM2026.05 | 0.91 | 0.23 | 1.14 | 6.1 | |
| GLOBE-CEClassifier Model=XGBoost2026.05 | 0.92 | 0.25 | 1.16 | 5.5 |