Mathematics
2024, 12, 571
20 of 26
Figure 12.
Model accuracy and loss.
The classification report reveals the ensemble model’s nuanced understanding of
diverse attack vectors (Table
6
). With perfect precision and recall in the “Normal” and
“DDoS_UDP” categories, the model demonstrates its ability to identify clear-cut patterns of
network behavior and straightforward intrusion attempts. However, the “SQL_injection”
and “XSS” categories, with lower precision, indicate a propensity for false positives, where
the model’s judgment may benefit from additional refinement. The “Password” and
“Fingerprinting” categories, which show disparities in precision and recall, suggest a more
complex interplay of features that could require advanced analytical strategies to improve
classification accuracy.