Time-Series Anomaly Detection on IOPS
69.3PrecisionTimesFM
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| TimesFMpretraining=large-scale2026.02 | 69.3 | 11.5 | 16.2 | 26.4 | 30.8 | |
| AnomaMindbackbone=Qwen3-8B, agent_framework=agent-lightning, optimization=full fine-tuning2026.02 | 63.5 | 44.6 | 46.7 | 46.8 | 50.4 | |
| Chronospretraining=large-scale2026.02 | 61.5 | 11.1 | 15 | 25.4 | 28.2 | |
| LLMAD2026.02 | 60.1 | 18 | 27.7 | 23.4 | 32.3 | |
| FFT-AD2026.02 | 51.1 | 18.6 | 22 | 36.8 | 32.1 | |
| CNN2026.02 | 50.2 | 18.6 | 21.8 | 30.9 | 30.4 | |
| LSTMAD2026.02 | 49.2 | 17.8 | 18.6 | 33.7 | 29.8 | |
| M2N22026.02 | 43.5 | 15.1 | 17.1 | 32.3 | 27 | |
| SR2026.02 | 31.7 | 25.6 | 23.7 | 29.4 | 27.6 | |
| TransAD2026.02 | 21.6 | 10.7 | 10.1 | 39.9 | 20.6 | |
| OFA2026.02 | 14.5 | 19.1 | 13.6 | 27.4 | 18.6 | |
| LLM-TSAD2026.02 | 12.9 | 42.7 | 15.7 | 15.7 | 21.8 |