IoT Intrusion Detection on N-BaIoT Scenario 2
99.01AccuracySecureDyn-FL
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SecureDyn-FLAttack Type=Same Model, Scenario=N/A2026.01 | 99.01 | 98.93 | 0.0405 | |
| FedAvgAttack Type=None, Scenario=N/A2026.01 | 98.89 | 98.72 | — | |
| FL TrustAttack Type=Same Model, Scenario=N/A2026.01 | 98.62 | 98.01 | 0.0097 | |
| FedAvgAttack Type=Label-Flipping, Scenario=Benign2026.01 | 98.57 | 98.85 | 0.0241 | |
| FL TrustAttack Type=None, Scenario=N/A2026.01 | 98.45 | 98.54 | — | |
| SecureDyn-FLAttack Type=None, Scenario=N/A2026.01 | 98.4 | 98.38 | — | |
| SecureDyn-FLAttack Type=Model Scaling, Scenario=N/A2026.01 | 98.23 | 97.61 | 0.0341 | |
| Shield FLAttack Type=None, Scenario=N/A2026.01 | 98.19 | 98.31 | — | |
| SecureDyn-FLAttack Type=Label-Flipping, Scenario=Attack2026.01 | 98.14 | 97.75 | 0.0738 | |
| FL-DefenderAttack Type=None, Scenario=N/A2026.01 | 98.03 | 98.7 | — | |
| Trimmed MeanAttack Type=None, Scenario=N/A2026.01 | 97.54 | 97.68 | — | |
| Shield FLAttack Type=Label-Flipping, Scenario=Both2026.01 | 97.49 | 97.12 | 0.0339 | |
| SecureDyn-FLAttack Type=Label-Flipping, Scenario=Both2026.01 | 96.94 | 97.62 | 0.0421 | |
| FL-DefenderAttack Type=Model Scaling, Scenario=N/A2026.01 | 96.57 | 97.02 | 0.0291 | |
| SecureDyn-FLAttack Type=Label-Flipping, Scenario=Benign2026.01 | 96.01 | 96.81 | 0.0061 | |
| Shield FLAttack Type=Label-Flipping, Scenario=Attack2026.01 | 94.51 | 92.55 | 0.1351 | |
| Shield FLAttack Type=Same Model, Scenario=N/A2026.01 | 94.41 | 94.75 | 0.0432 | |
| FedAvgAttack Type=Label-Flipping, Scenario=Attack2026.01 | 92.61 | 91.53 | 0.1365 | |
| FedAvgAttack Type=Label-Flipping, Scenario=Both2026.01 | 83.85 | 82.65 | 0.1558 | |
| FL TrustAttack Type=Label-Flipping, Scenario=Attack2026.01 | 75.05 | 63.75 | 0.613 | |
| Trimmed MeanAttack Type=Same Model, Scenario=N/A2026.01 | 72.54 | 62.2 | 0.2885 | |
| FL-DefenderAttack Type=Label-Flipping, Scenario=Attack2026.01 | 53.14 | 0 | 1 | |
| Trimmed MeanAttack Type=Label-Flipping, Scenario=Attack2026.01 | 53.05 | 0 | 1 | |
| FL-DefenderAttack Type=Same Model, Scenario=N/A2026.01 | 51.63 | 0 | 0.4453 | |
| Trimmed MeanAttack Type=Label-Flipping, Scenario=Both2026.01 | 47.98 | 63.61 | 0.5163 | |
| FL TrustAttack Type=Label-Flipping, Scenario=Both2026.01 | 46.95 | 63.42 | 0.5169 | |
| FL-DefenderAttack Type=Label-Flipping, Scenario=Both2026.01 | 46.81 | 63.78 | 0.5181 | |
| Shield FLAttack Type=Label-Flipping, Scenario=Benign2026.01 | 46.41 | 62.95 | 1 | |
| FL-DefenderAttack Type=Label-Flipping, Scenario=Benign2026.01 | 46.41 | 63.4 | 1 | |
| FL TrustAttack Type=Label-Flipping, Scenario=Benign2026.01 | 46.39 | 63.54 | 1 | |
| FedAvgAttack Type=Model Scaling, Scenario=N/A2026.01 | 46.39 | 63.73 | 0.5089 | |
| Shield FLAttack Type=Model Scaling, Scenario=N/A2026.01 | 46.38 | 63.25 | 0.5241 | |
| FL TrustAttack Type=Model Scaling, Scenario=N/A2026.01 | 46.37 | 63.47 | 0.5169 | |
| Trimmed MeanAttack Type=Label-Flipping, Scenario=Benign2026.01 | 46.36 | 63.41 | 1 | |
| FedAvgAttack Type=Same Model, Scenario=N/A2026.01 | 46.01 | 62.91 | 0.5267 | |
| Trimmed MeanAttack Type=Model Scaling, Scenario=N/A2026.01 | 45.95 | 62.91 | 0.5163 |