Table 6 Comparative performance evaluation of fault diagnosis methods across varying sample sizes.
From: RS-SCBiGRU: a noise-robust neural network for high-speed motor fault diagnosis with limited samples
Methods | 10 samples | 20 samples | 30 samples | 40 samples | ||||
|---|---|---|---|---|---|---|---|---|
Accuracy (%) | F1 score (%) | Accuracy (%) | F1 score (%) | Accuracy (%) | F1 score (%) | Accuracy (%) | F1 score (%) | |
RS-SCBiGRU | 84.02 | 82.78 | 92.55 | 92.37 | 90.35 | 90.20 | 93.23 | 93.25 |
SCBiGRU | 82.84 | 82.20 | 91.36 | 91.16 | 89.89 | 89.66 | 90.40 | 90.54 |
BiGRU | 74.18 | 72.55 | 81.33 | 81.58 | 81.99 | 82.01 | 84.02 | 83.71 |
BiLSTM | 73.88 | 72.13 | 78.05 | 77.82 | 82.21 | 82.94 | 85.96 | 85.59 |
DAMN | 78.62 | 77.23 | 81.29 | 80.34 | 83.09 | 82.99 | 85.77 | 85.61 |
MSCNN | 82.65 | 82.93 | 91.90 | 91.52 | 80.91 | 79.13 | 81.49 | 79.60 |
MSCNN-LSTM | 38.32 | 37.36 | 56.90 | 55.09 | 73.95 | 72.00 | 82.40 | 80.86 |
RNN-WDCNN | 74.51 | 73.30 | 86.28 | 86.07 | 86.82 | 86.52 | 90.16 | 89.85 |
WDCNN | 58.75 | 57.12 | 76.85 | 75.62 | 86.30 | 85.96 | 88.72 | 88.63 |
