Table 8 Grid search results of the classifiers for module 2, with and without L 0 penalization
Classifier | Linear reg | Lasso | Ridge | Elastic net | Relaxed Lasso | L1 logreg | L2 logreg | LDA |
|---|---|---|---|---|---|---|---|---|
L 0 penalized ROC AUC | 73.7 | 84.1 | 75.4 | 84.6 | 84.3 | 83.3 | 75.4 | 86.5 |
Associated real ROC AUC | 80.0 | 84.9 | 81.5 | 85.5 | 85.1 | 84.2 | 85.4 | 87.4 |
Features used with L 0 | 36.5 | 4.4 | 35.7 | 5.3 | 4.4 | 5.4 | 58.0 | 5.0 |
Not penalized ROC AUC | 79.3 | 85.1 | 81.2 | 86.3 | 85.3 | 85.9 | 85.3 | 87.4 |
Features used without L 0 | 35.4 | 4.5 | 34.6 | 10.7 | 4.5 | 4.0 | 58.0 | 5.0 |
Classifier | pSVM | rSVM | eSVM | L1 lSVM | Grad Boost | AdaBoost | Rand Forest | Tree |
L 0 penalized ROC AUC | 86.8 | 86.1 | 49.3 | 81.2 | 85.7 | 85.7 | 86.3 | 82.4 |
Associated real ROC AUC | 87.7 | 86.6 | 0.50 | 81.6 | 86.6 | 86.6 | 86.2 | 83.0 |
Features used with L 0 | 5.0 | 5.0 | 4.0 | 2.4 | 5.0 | 5.0 | 5.0 | 3.0 |
Not penalized ROC AUC | 87.6 | 85.9 | 50.0 | 83.0 | 86.6 | 86.6 | 88.4 | 83.0 |
Features used without L 0 | 5.0 | 12.0 | 4.0 | 29.4 | 5.0 | 5.0 | 28.9 | 3.0 |