Table 3 Predictive performance comparisons with different learning methods in Independent Dataset (Dataset 3).
From: Improved Classification of Blood-Brain-Barrier Drugs Using Deep Learning
Prediction performance in dataset 3 | Method | Sigmoid-SVM | Poly-SVM | RBF-SVM | ||||||
Side Effects(SE) | Indications | SE + Indications | Side Effects(SE) | Indications | SE + Indications | Side Effects(SE) | Indications | SE + Indications | ||
| Â | Prediction | Prediction | Prediction | Prediction | Prediction | Prediction | Prediction | Prediction | Prediction | |
AUC | 0.675 ± 0.15 | 0.72 ± 0.18 | 0.672 ± 0.16 | 0.7058 ± 0.19 | 0.71 ± 0.15 | 0.685 ± 0.12 | 0.71 ± 0.12 | 0.71 ± 0.16 | 0.69 ± 0.15 | |
Accuracy | 0.51 ± 0.05 | 0.64 ± 0.18 | 0.64 ± 0.02 | 0.671 ± 0.14 | 0.53 ± 0.13 | 0.54 ± 0.11 | 0.735 ± 0.18 | 0.63 ± 0.12 | 0.69 ± 0.11 | |
F1 | 0.42 ± 0.06 | 0.542 ± 0.21 | 0.584 ± 0.09 | 0.521 ± 0.16 | 0.51 ± 0.2 | 0.41 ± 0.06 | 0.601 ± 0.21 | 0.62 ± 0.15 | 0.561 ± 0.16 | |
Method | KNN | Decision Tree (DT) | Deep Learning | |||||||
Side Effects(SE) | Indications | SE + Indications | Side Effects(SE) | Indications | SE + Indications | Side Effects(SE) | Indications | SE + Indications | ||
| Â | Prediction | Prediction | Prediction | Prediction | Prediction | Prediction | Prediction | Prediction | Prediction | |
AUC | 0.73 ± 0.03 | 0.71 ± 0.11 | 0.734 ± 0.16 | 0.568 ± 0.16 | 0.67 ± 0.18 | 0.568 ± 0.12 | 0.978 ± 0.02 | 0.98 ± 0.02 | 0.99 ± 0.01 | |
Accuracy | 0.7 ± 0.11 | 0.72 ± 0.15 | 0.74 ± 0.19 | 0.584 ± 0.2 | 0.66 ± 0.18 | 0.59 ± 0.18 | 0.978 ± 0.02 | 0.964 ± 0.03 | 0.98 ± 0.02 | |
F1 | 0.73 ± 0.07 | 0.71 ± 0.11 | 0.692 ± 0.11 | 0.516 ± 0.13 | 0.64 ± 0.16 | 0.519 ± 0.15 | 0.91 ± 0.09 | 0.896 ± 0.1 | 0.918 ± 0.08 | |
