Table 4 Comparative performance of different methods on AWID3 dataset for KRACK and Kr00k multiclass detection.
Year | Method | Accuracy | Precision | Recall (TPR) | F1-score | FPR |
|---|---|---|---|---|---|---|
2023 | ANOVA algorithm technique51 | 0.90 | 0.88 | 0.89 | 0.88 | 0.07 |
2021 | Unsupervised framework50 | 0.88 | 0.84 | 0.86 | 0.85 | 0.09 |
2022 | Best of both worlds52 | 0.93 | 0.91 | 0.92 | 0.91 | 0.05 |
2022 | Multi-class neural network53 | 0.93 | 0.90 | 0.91 | 0.90 | 0.06 |
2022 | kTRACKER49 | 0.96 | 0.94 | 0.95 | 0.94 | 0.05 |
2024 | SM-GBT54 | 0.97 | 0.95 | 0.96 | 0.95 | 0.04 |
2024 | AttackNet (CNN-GRU)38 | 0.96 | 0.97 | 0.94 | 0.96 | 0.03 |
2025 | Cyber-Sentinet (ResNet + SHAP)40 | 0.97 | 0.98 | 0.96 | 0.99 | 0.02 |
2025 | CPS-IIoT-P2Attention41 | 0.97 | 0.98 | 0.96 | 0.97 | 0.03 |
2025 | CWFLAM-VAE (XGBoost)44 | 0.96 | 0.97 | 0.91 | 0.94 | 0.03 |
2025 | ML Model Pipeline 2 (Ours) | 0.98 | 0.98 | 0.98 | 0.98 | 0.02 |
