Federated Learning Classification on Edge-IoT
93.2Label Flip AccuracyZTA-FL
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ZTA-FLByzantine Attacker Rate=30%2025.12 | 93.2 | 91.7 | 8.7 | |
| FLAMEByzantine Attacker Rate=30%2025.12 | 90.1 | 89.7 | 12.8 | |
| FLTrustByzantine Attacker Rate=30%2025.12 | 89.4 | 88.2 | 15.3 | |
| RFAByzantine Attacker Rate=30%2025.12 | 87.6 | 86.9 | 22.4 | |
| Trimmed MeanByzantine Attacker Rate=30%2025.12 | 85.1 | 84.6 | 38.7 | |
| KrumByzantine Attacker Rate=30%2025.12 | 82.4 | 83.9 | 45.2 | |
| FedAvgByzantine Attacker Rate=30%2025.12 | 67.8 | 71.3 | 87.3 |