Table 2.
Prediction performance (area under the receiver operatic characteristic curve) for combinations of features and classification methods.
| Classifiers | Feature generation algorithm | ||
|
|
Bag of features | Node2vec | Bag of features+Node2vec |
|
|
AUROCa (%) | AUROC (%) | AUROC (%) |
| Random forest | 94.82 | 91.89 | 96.19 |
| Naive Bayes | 92.30 | 92.91 | 94.76 |
| Logistic regression | 86.68 | 85.25 | 89.39 |
| Support vector machine | 84.62 | 83.92 | 86.72 |
| Convolutional neural network | 64.14 | 63.36 | 57.68 |
| Deep neural network | 92.56 | 92.87 | 95.12 |
| Graph convolutional networks | 79.67 | 83.62 | 83.83 |
aAUROC: area under the receiver operating characteristic curve.