Research Article
A Diagnosis Framework for High-reliability Equipment with Small Sample Based on Transfer Learning
Table 3
Accuracy of diagnosis under different transfer learning approaches.
| | Model | Accuracy |
| | Proposed model | 87.3±0.5% | | ResNet-50 (He et al., 2016) | 63.2±0.8% | | DANN (Ganin et al., 2016) | 80.1±0.2% | | ADDA (Tzeng et al., 2017) | 53.8±1.5% | | JAN (Long et al., 2017) | 82.3±0.6% | | MADA (Pei et al., 2018) | 76.3±1.2% | | CBST (Zou et al., 2018) | 56.7±0.9% | | CAN (Zhang et al., 2018) | 75.1±0.5% | | CDAN+E (Long et al., 2018b) | 79.5±0.7% | | DM-ADA (Xu et al., 2020) | 58.7±1.1% | | 3CATN (Li et al., 2019) | 83.2±0.6% | | ALDA (Chen et al., 2020) | 73.6±0.5% |
|
|
