TABLE 3.
Repeated features across different partitions for the first run of the GA (“Common features”, see Figures 2, 3) using different ML models. The frequency each feature is repeated within partitions is shown in paretheses. The F1/2 Train and F1/2 Test columns refer to the average score across the different training folds and test folds, respectively.
| Model | Feature 1 | Feature 2 | Feature 3 | Feature 4 | Feature 5 | Feature 6 | F1/2 train | F1/2 test |
|---|---|---|---|---|---|---|---|---|
| RF | KRFPC476 (2) | KRFPC3707 (2) | KRFPC4556 (2) | SubFPC3 (2) | SubFPC301 (2) | MP (2) | 1.0 | - |
| LDA | KRFPC413 (2) | KRFPC566 (3) | KRFPC608 (4) | KRFPC4830 (2) | - | - | 0.80 | 0.77 |
| SVM | KRFPC1564 (2) | KRFPC3946 (2) | SubFPC169 (2) | - | - | 0.72 | - | |
| LR | KRFPC608 (4) | ROTB (2) | - | - | - | - | 0.59 | 0.54 |