Table 4 Comparative experiments with other models.
From: Elevator fault diagnosis based on digital twin and PINNs-e-RGCN
Model | Maximum accuracy (%) | Minimum accuracy (%) | Average accuracy (%) |
|---|---|---|---|
e-RGCN | 96.88 | 96.31 | 96.61± 0.17 |
Graph SAGE | 94.66 | 93.74 | 94.18± 0.32 |
SGCN | 93.51 | 92.72 | 93.37± 0.48 |
GAT | 93.30 | 92.19 | 92.70± 0.42 |
DGCN | 91.89 | 90.68 | 91.44± 0.46 |
LeNet-5 | 89.67 | 88.33 | 88.93± 0.43 |
Res Net | 87.23 | 86.38 | 86.75±0.30 |
