Research Article
Toward Identifying APT Malware through API System Calls
Table 2
Binary classification evaluation results of classical machine learning algorithms.
| Classification algorithm | Accuracy | Precision | Recall | F1 |
| KNN | 0.9455 | 0.9268 | 0.8333 | 0.8776 | Logistic regression | 0.9214 | 0.8097 | 0.8026 | 0.8062 | Decision tree | 0.9679 | 0.9324 | 0.9079 | 0.92 | Gradient boosting | 0.9214 | 0.986 | 0.9298 | 0.9571 | AdaBoost | 0.9282 | 0.9061 | 0.7193 | 0.802 | Gaussian NB | 0.8045 | 0.6364 | 0.2456 | 0.3544 | Linear discriminant analysis | 0.9179 | 0.925 | 0.6491 | 0.7629 | Quadratic discriminant | 0.9161 | 0.9653 | 0.6096 | 0.7473 | SVC | 0.9473 | 0.9617 | 0.7719 | 0.8564 | Multinomial NB | 0.3187 | 0.2102 | 0.8684 | 0.3385 |
|
|