Anomaly Detection on SWaT (P/R/F1 and eTa Metrics)
98.1Precision (Pre)MTAD-GAT
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MTAD-GATCategory=Graph Learning-based Methods2026.06 | 98.1 | 75.3 | 85.2 | 73.4 | 60.9 | 66.6 | |
| USADCategory=Sequence Representation Learning-based Methods2026.06 | 97.9 | 67.7 | 80 | 77.4 | 52.3 | 62.4 | |
| IstGPT2026.06 | 97.7 | 92.6 | 95.1 | 92.2 | 81.7 | 86.7 | |
| GDNCategory=Graph Learning-based Methods2026.06 | 96.4 | 81.1 | 88.1 | 83.3 | 76.5 | 79.8 | |
| ICS-ADCategory=Sequence Representation Learning-based Methods2026.06 | 95.6 | 85.2 | 90.1 | 81.8 | 70.2 | 75.6 | |
| MSCREDCategory=Sequence Representation Learning-based Methods2026.06 | 95.5 | 52.4 | 67.7 | 69.8 | 38.6 | 49.7 | |
| GTACategory=Graph Learning-based Methods2026.06 | 94.8 | 89.4 | 92 | 82.5 | 85.7 | 84.1 | |
| GeCosCategory=Traditional Rule-based Methods2026.06 | 94.8 | 79 | 86.2 | 83.1 | 60.7 | 70.2 | |
| VAE-LSTMCategory=Sequence Representation Learning-based Methods2026.06 | 89.5 | 65.8 | 75.8 | 58.7 | 40.3 | 47.8 | |
| OmniAnomalyCategory=Sequence Representation Learning-based Methods2026.06 | 88 | 72.3 | 79.4 | 82.9 | 77.2 | 79.9 | |
| JumpStarterCategory=Traditional Rule-based Methods2026.06 | 86.1 | 79.7 | 82.8 | 82.9 | 78.5 | 80.6 | |
| LLMTIMECategory=LLMs-based Methods2026.06 | 58.9 | 60.6 | 59.7 | 76.2 | 55.5 | 64.2 | |
| SigllmCategory=LLMs-based Methods2026.06 | 53.1 | 74.7 | 62.1 | 48.4 | 61.6 | 54.2 |