Research Article
Deep Multiscale Soft-Threshold Support Vector Data Description for Enhanced Heavy-Duty Gas Turbine Generator Sets’ Anomaly Detection
Table 6
Computational resources consumed by model training.
| Model | The initial stage | The second stage | Computation/FLOPs (M) | Number of parameters (106) | Computation/FLOPs (M) | Number of parameters (106) |
| Deep-SVDD | 92 | 8.0 | 72 | 6.6 | Deep-SVDD + S | 142 | 10.3 | 102 | 8.4 | Deep-SVDD + M | 264 | 14.7 | 183 | 11.4 | DMS-SVDD | 292 | 16.3 | 215 | 13.9 |
|
|