Table 2.
Modela | Unpenalised | Penalised | Exclusion |
---|---|---|---|
Lasso+sel (cv) | 0.72 | 0.73 | 0.69 |
Lasso+sel (10) | 0.72 | 0.72 | 0.67 |
Lasso+sel (50) | 0.73 | 0.73 | 0.70 |
ridge | 0.72 | 0.69 | 0.68 |
GRridge | 0.73 | 0.72 | 0.69 |
GRridge+sel (lasso) | 0.72 | 0.76 | 0.72 |
GRridge+sel (10) | 0.72 | 0.73 | 0.71 |
GRridge+sel (50) | 0.72 | 0.75 | 0.72 |
GRlasso+sel (cv) | 0.73 | 0.73 | 0.67 |
GRlasso+sel (10) | 0.72 | 0.73 | 0.70 |
GRlasso+sel (50) | 0.72 | 0.72 | 0.68 |
MR | 0.74 | 0.73 | 0.71 |
RF | 0.71 | 0.67 | 0.67 |
CoRF | 0.73 | 0.63 | 0.59 |
aNumbers in parentheses indicate the number of features selected in the model, that is, optimal number (cv) of features, 10 features or 50 features
cv, cross-validation; MR, multiridge; RF, random forest; sel, selection