Table 3 Experimental accuracy at different sample sizes.
From: Elevator fault diagnosis based on digital twin and PINNs-e-RGCN
Sample size | Real label | Predicted label | Average accuracy | |||
|---|---|---|---|---|---|---|
Normal | Jump | Star-stop | Acuteness | |||
120 | Normal | 0.95 | 0 | 0.05 | 0 | 93.25% |
Jump | 0 | 0.88 | 0.02 | 0.10 | Â | |
Star-stop | 0.06 | 0 | 0.92 | 0.02 | Â | |
Acuteness | 0 | 0.02 | 0 | 0.98 | Â | |
160 | Normal | 1 | 0 | 0 | 0 | 96.5% |
Jump | 0 | 0.96 | 0 | 0.04 | Â | |
Star-stop | 0.05 | 0 | 0.95 | 0 | Â | |
Acuteness | 0 | 0 | 0 | 1 | Â | |
200 | Normal | 1 | 0 | 0 | 0 | 96.8% |
Jump | 0 | 0.97 | 0 | 0.03 | Â | |
Star-stop | 0.04 | 0 | 0.96 | 0 | Â | |
Acuteness | 0 | 0 | 0 | 1 | Â | |
