Table 4.
Average Success Index calculated using a 10-fold cross-validation
| Dataset | RGIFE-Min | RGIFE-Max | RGIFE-Union | CFS | ReliefF | SVM-RFE | Chi-Square | L1 |
|---|---|---|---|---|---|---|---|---|
| CorrAL | 59.93 | 77.11 | 87.07 | 84.57 | 72.04 | 64.53 | 82.06 | 84.11 |
| XOR-100 | 89.99 | 79.88 | 89.88 | 24.72 | 49.86 | 14.84 | 9.84 | 79.30 |
| Parity3+3 | 44.44 | 44.44 | 76.67 | -5.93 | 76.67 | -15.93 | -8.52 | 5.56 |
| Monk3 | 84.17 | 84.17 | 84.17 | 62.50 | 84.17 | N/A | 59.17 | 73.33 |
| Madelon | 59.98 | 77.98 | 87.97 | 17.99 | 89.97 | 23.97 | 39.97 | 99.01 |
| Average | 67.70 | 72.72 | 85.15 | 36.77 | 74.54 | 21.85 | 36.50 | 68.26 |
The last row reports the average values across the five datasets. The highest indexes are shown in bold. N/A is used for SVM-RFE when analysing the Monk3 dataset as the method cannot deal with categorical attributes