Table 3 Adversarial attack.
Model Variant | Attack Type | Detection Accuracy (%) | False Positive Rate (%) |
|---|---|---|---|
M-LDAE (Baseline) | Normal DDoS | 96.8 | 3.5 |
M-LDAE (Adversarial Training) | FGSM-DDoS | 92.4 | 5.1 |
M-LDAE (Adversarial Training) | PGD-MITM | 89.7 | 6.8 |
Ensemble (M-LDAE + GNN + XGBoost) | PGD-MITM | 94.5 | 4.2 |
