Table 3
Summary of performance benchmarks of comparison methods.
| Method | Input signals | Feature extraction | Model architecture | Supports Multi-fault decoupling |
|---|---|---|---|---|
| SVM Multi-feature fusion | Current, vibration, displacement | 32-dimensional fused handcrafted features | Support vector machine | Limited (single-label output) |
| ResNet Time-series model | Concatenated 4-channel raw signals | End-to-end learning | 1D ResNet-18 | No |
| Proposed ethod (Full) | 4 modalities + Graph embedding | Time-frequency + GNN + Dual attention | Multimodal fusion network | Yes |
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.