Research Article
A VNS-EDA Algorithm-Based Feature Selection for Credit Risk Classification
Table 3
Performance metrics of VNS-EDA with the optimal feature subset.
| Dataset | Accuracy (%) | F-measure (%) | Sensitivity (%) | Specificity (%) | Training | Testing | Training | Testing | Training | Testing | Training | Testing |
| ZhongAn Credit | 86.3 | 85.6 | 77.2 | 76.3 | 61.5 | 60.1 | 91.2 | 90.6 | German Bank | 76.2 | 74.5 | 72.9 | 71.4 | 65.6 | 63.8 | 87.7 | 85.1 | LendingClub | 74.3 | 75.2 | 70.6 | 71.8 | 60.2 | 61.5 | 87.9 | 89.3 |
|
|