Table 1 Test accuracy using CNN24 model based on IID and Sharding.
From: Efficient federated learning for pediatric pneumonia on chest X-ray classification
Algorithms | IID | Sharding = 2 | Sharding = 4 | Sharding = 8 | Sharding = 16 | Sharding = 32 |
|---|---|---|---|---|---|---|
FedDDA19 | 97.64% | 93.86% | 95.34% | 97.27% | 97.34% | 97.55% |
FedProx20 | 97.13% | 95.61% | 95.54% | 96.67% | 97.05% | 96.86% |
FedProx+DM | 97.40% | 95.28% | 95.74% | 97.18% | 97.26% | 97.17% |
FedProxM | 97.71% | 97.37% | 97.85% | 98.39% | 97.52% | 98.63% |
SCAFFOLD21 | 97.27% | 93.46% | 97.01% | 94.81% | 94.16% | 94.84% |
SCAFFOLD+DM | 98.07% | 13.52% | 96.81% | 97.02% | 13.52% | 13.52% |
SCAFFOLDM | 98.49% | 99.27% | 98.29% | 98.24% | 98.48% | 99.06% |
FedCurv22 | 94.88% | 74.97% | 94.25% | 94.30% | 94.55% | 94.18% |
FedCurv+DM | 92.45% | 88.31% | 88.82% | 91.44% | 92.98% | 93.26% |
FedCurvM | 95.97% | 85.83% | 94.28% | 93.66% | 94.96% | 95.49% |
FedNTD23 | 97.18% | 96.00% | 96.17% | 96.70% | 96.77% | 96.98% |
FedNTD+DM | 97.07% | 93.26% | 94.45% | 96.84% | 96.53% | 96.87% |
FedNTDM | 98.39% | 99.44% | 97.13% | 98.60% | 97.98% | 98.63% |
