Advances, Systems and Applications
From: Bearing fault diagnosis method based on improved Siamese neural network with small sample
Dataset | Testing set 2 | Testing set 3 | ||||||
---|---|---|---|---|---|---|---|---|
1D CNN | Prototype net | Siamese net | ISNN | 1D CNN | Prototype net | Siamese net | ISNN | |
Training set A | 49.20% | 42.30% | 46.40% | 83.00% | 47.30% | 40.40% | 45.10% | 82.40% |
Training set B | 67.00% | 54.40% | 58.30% | 89.70% | 65.20% | 51.60% | 55.20% | 88.70% |
Training set C | 85.60% | 65.80% | 69.60% | 94.80% | 83.50% | 60.90% | 66.00% | 94.40% |
Training set D | 90.20% | 74.40% | 78.50% | 97.80% | 86.00% | 68.60% | 76.60% | 95.20% |