Table 5 Selected features and category composition by method (combined features).
From: Consensus-driven feature selection for transparent and robust loan default prediction
Dataset combine | Techniques | Features | Selected | Category Distribution (Selected/Total) |
|---|---|---|---|---|
Firms | Stepwise XGB | 84 | 24 | 16/47, 0/10, 2/10, 6/8, 0/9, |
| Â | MI XGB | 84 | 23 | 18/47, 2/10, 0/10, 0/8, 3/9, |
| Â | ReliefF XGB | 84 | 26 | 16/47, 0/10, 7/10, 2/8, 1/9, |
| Â | ElasticNet XGB | 84 | 22 | 13/47, 2/10, 4/10, 1/8, 2/9, |
| Â | RFE XGB | 84 | 25 | 17/47, 1/10, 4/10, 2/8, 1/9, |
|  | HRA‑FS XGB | 84 | 23 | 12/47, 2/10, 3/10, 2/8, 4/9, |
Farmers | Stepwise XGB | 46 | 21 | 8/12, 8/17, 1/6, 2/4, 2/7, |
| Â | MI XGB | 46 | 23 | 7/12, 2/17, 5/6, 3/4, 6/7, |
| Â | ReliefF XGB | 46 | 20 | 9/12, 4/17, 2/6, 0/4, 5/7, |
| Â | ElasticNet XGB | 46 | 24 | 10/12, 3/17, 4/6, 2/4, 5/7, |
| Â | RFE XGB | 46 | 22 | 10/12, 2/17, 3/6, 1/4, 6/7, |
|  | HRA‑FS XGB | 46 | 21 | 7/12, 8/17, 3/6, 1/4, 2/7, |
