Skip to main content

Table 2 SNF-NN performance comparison with baseline machine learning methods on the SND benchmark dataset

From: SNF-NN: computational method to predict drug-disease interactions using similarity network fusion and neural networks

Method Accuracy Specificity Precision Recall F1-score MCC AUC-ROC AUC-PR
SAMME 0.682 0.841 0.767 0.523 0.621 0.384 0.682 0.764
DT 0.553 0.985 0.888 0.122 0.214 0.211 0.553 0.725
GPC 0.646 0.676 0.655 0.615 0.634 0.292 0.646 0.731
KNN 0.650 0.619 0.641 0.681 0.661 0.300 0.650 0.741
GNB 0.669 0.586 0.645 0.751 0.694 0.342 0.669 0.760
QDA 0.649 0.632 0.646 0.666 0.654 0.300 0.649 0.740
RF 0.533 0.692 0.635 0.374 0.354 0.112 0.533 0.661
Linear-SVM 0.702 0.718 0.709 0.685 0.697 0.404 0.702 0.776
RBF-SVM 0.535 0.949 0.704 0.120 0.204 0.124 0.535 0.632
SNF-NN 0.796 0.777 0.785 0.816 0.800 0.593 0.867 0.876
  1. The best value of each evaluation metric is shown in bold