Research Article
Structural Vibration Data Anomaly Detection Based on Multiple Feature Information Using CNN-LSTM Model
Table 3
Performance indexes of the LSTM-CNN model on the validation set.
| Class | Anomalous patterns | Precision (%) | Recall (%) | F1 score (%) |
| 1 | Normal | 97.33 | 98.06 | 97.69 | 2 | Outlier | 96.83 | 96.30 | 96.56 | 3 | Square | 98.63 | 99.81 | 99.22 | 4 | Missing | 99.80 | 96.20 | 97.97 | 5 | Minor | 96.83 | 99.00 | 97.90 | Accuracy | 97.87% |
|
|