Advances, Systems and Applications
Table 3 Accuracy of different feature extraction methods
From: Bearing fault diagnosis method based on improved Siamese neural network with small sample
Feature extraction method | Training A | Training B | Training C | Training D |
|---|---|---|---|---|
ISNN | 84.10% | 92.30% | 95.00% | 98.10% |
Time domain | 43.40% | 48.40% | 55.60% | 63.20% |
Frequency domain | 80.20% | 87..8% | 92.80% | 95.20% |
Fully connected | 77.40% | 84.60% | 88.00% | 93.80% |
1D CNN | 72.60% | 82.40% | 91.60% | 97.20% |