Research Article
Session-Based Webshell Detection Using Machine Learning in Web Logs
Table 2
The influence of different thresholds in session identification.
| Threshold (%) | Accuracy | Recall | Precision | F1 score |
| 55 | 0.932 | 0.8624 | 0.8757 | 0.8690 | 60 | 0.9364 | 0.869 | 0.8958 | 0.8822 | 65 | 0.9598 | 0.9568 | 0.8977 | 0.9263 | 70 | 0.9597 | 0.9615 | 0.9036 | 0.9317 | 75 | 0.9696 | 0.9368 | 0.9009 | 0.9185 | 80 | 0.9376 | 0.9401 | 0.8518 | 0.8937 | 85 | 0.8974 | 0.7228 | 0.6932 | 0.7077 | 90 | 0.8779 | 0.7186 | 0.6418 | 0.6780 |
|
|