Research Article
MTAD-TF: Multivariate Time Series Anomaly Detection Using the Combination of Temporal Pattern and Feature Pattern
Table 2
Performance of our model and baselines.
| Dataset | MSL | SMAP | SMD | Precision | Recall | F1 | Precision | Recall | F1 | Precision | Recall | F1 |
| LSTM-NDT | 0.5944 | 0.5374 | 0.5640 | 0.8965 | 0.8846 | 0.8905 | 0.5684 | 0.6438 | 0.6037 | LSTM-VAE | 0.5257 | 0.9546 | 0.6780 | 0.8551 | 0.6366 | 0.7298 | 0.7922 | 0.7075 | 0.7842 | DAGMM | 0.5412 | 0.9934 | 0.7007 | 0.5845 | 0.9058 | 0.7105 | 0.5835 | 0.9042 | 0.7093 | OmniAnomaly | 0.8867 | 0.9117 | 0.8989 | 0.7416 | 0.9776 | 0.8434 | 0.8334 | 0.9449 | 0.8857 |
| MTAD-TF | 0.9043 | 0.8988 | 0.9015 | 0.9779 | 0.8192 | 0.8916 | 0.9045 | 0.9048 | 0.8940 |
|
|