Table 3 Test accuracy using CNN model based on S_Max and Centralized.
From: Efficient federated learning for pediatric pneumonia on chest X-ray classification
Algorithms | S_Max=4 | S_Max=10 | S_Max=20 | S_Max=50 | Centralized |
|---|---|---|---|---|---|
FedDDA | 95.27% | 95.30% | 93.95% | 93.74% | 97.51% |
FedProx | 95.03% | 95.71% | 95.66% | 95.87% | 96.84% |
FedProx+DM | 94.22% | 94.45% | 93.94% | 93.75% | 97.27% |
FedProxM | 97.44% | 97.61% | 97.22% | 97.20% | 98.63% |
SCAFFOLD | 94.71% | 95.59% | 95.41% | 95.56% | 96.57% |
SCAFFOLD+DM | 94.99% | 92.99% | 94.49% | 94.87% | 97.99% |
SCAFFOLDM | 97.95% | 97.06% | 96.96% | 96.19% | 98.50% |
FedCurv | 72.97% | 85.02% | 91.58% | 75.99% | 94.71% |
FedCurv+DM | 90.35% | 86.48% | 86.48% | 86.48% | 93.36% |
FedCurvM | 80.89% | 91.03% | 93.03% | 90.30% | 94.89% |
FedNTD | 96.61% | 95.66% | 96.06% | 95.92% | 96.88% |
FedNTD+DM | 94.40% | 94.83% | 93.12% | 93.25% | 97.05% |
FedNTDM | 97.59% | 97.56% | 98.02% | 97.27% | 98.74% |
