Table 2.
The result of AUC (Area under Curve) and accuracy for models based on test set of 14 nucleoside derivatives
| Models | Features | Test set performance | ||||
|---|---|---|---|---|---|---|
| Accuracy | F1 Score | Precision | Recall | AUC | ||
| DT | Descriptor_ REF # | 0.60 | 0.67 | 0.75 | 0.60 | 0.60 |
| LR | Descriptor_ REF # | 0.67 | 0.76 | 1.00 | 0.61 | 0.81 |
| RF | Descriptor_ REF # | 0.53 | 0.59 | 0.63 | 0.56 | 0.53 |
| XGBoost | Descriptor_ REF # | 0.60 | 0.57 | 0.50 | 0.67 | 0.61 |
*LR Logistic regression, DT Decision tree, RF Random forest, XGBoost Extreme gradient boosting.
#Descriptors-REF: Recursive feature elimination (REF) has different optimal descriptors for different Algorithms: LR, n = 34; XGBoost, n = 33; DT, n = 23; RF, n = 26.