Open Access
Review
Table 4
Performance comparison and technical advantage analysis of transfer learning and hybrid strategy models.
| Method | Cross-domain accuracy retention rate (%) | Minimum sample requirement | Computational overhead (FLOPs) | Core advantages |
|---|---|---|---|---|
| MMD-TSM | 89.2 | 200 | 5.3 × 106 | Feature Distribution Alignment [31] |
| VA-WGAN | 93.5 | 150 | 8.7 × 106 | Sample Generation Enhancement [32] |
| AFARN | 98.0 | 250 | 11.2 × 106 | Physical Interpretability [12] |
| DL-MNPF | 95.8 | 300 | 15.4 × 106 | Multimodal Fusion [33] |
| MTST-CNN | 96.3 | 40 | 9.8 × 106 | Rapid Adaptation [18] |
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.