Probabilistic Marginal Forecasting on fev-bench marginal forecasting v0.7.0
91.7Win RateChronos-2
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Chronos-2Runtime (s)=0.8, Leak. (%)=0, # fails=02026.05 | 91.7 | 47.3 | |
| TiRexRuntime (s)=0.2, Leak. (%)=1, # fails=02026.05 | 83.4 | 42.6 | |
| TimesFM-2.5Runtime (s)=1.9, Leak. (%)=10, # fails=02026.05 | 78.6 | 42.2 | |
| TabPFN-TS-3Runtime (s)=234.6, Leak. (%)=0, # fails=02026.05 | 73.6 | 43.1 | |
| Toto-1.0Runtime (s)=22.1, Leak. (%)=8, # fails=02026.05 | 71.6 | 40.7 | |
| Moirai-2.0Runtime (s)=0.3, Leak. (%)=28, # fails=02026.05 | 66.2 | 39.3 | |
| Chronos-BoltRuntime (s)=0.2, Leak. (%)=0, # fails=02026.05 | 66.2 | 38.9 | |
| TabPFN-v2-TSRuntime (s)=88.9, Leak. (%)=0, # fails=22026.05 | 64.1 | 39.6 | |
| TabICL-v2Runtime (s)=64.7, Leak. (%)=0, # fails=02026.05 | 53.8 | 30.8 | |
| Sundial-BaseRuntime (s)=8.0, Leak. (%)=1, # fails=02026.05 | 47.1 | 33.4 | |
| Stat. EnsembleRuntime (s)=148.6, Leak. (%)=0, # fails=112026.05 | 43.8 | 20.2 | |
| AutoARIMARuntime (s)=19.5, Leak. (%)=0, # fails=102026.05 | 39.6 | 20.6 | |
| CatBoost (Recursive)Runtime (s)=0.3, Leak. (%)=0, # fails=02026.05 | 35.7 | 23 | |
| LightGBM (Recursive)Runtime (s)=0.3, Leak. (%)=0, # fails=02026.05 | 33.4 | 21.7 | |
| AutoETSRuntime (s)=3.5, Leak. (%)=0, # fails=32026.05 | 32.7 | -26.8 | |
| AutoThetaRuntime (s)=3.3, Leak. (%)=0, # fails=02026.05 | 27.1 | 5.5 | |
| Seasonal NaiveRuntime (s)=0.5, Leak. (%)=0, # fails=02026.05 | 19.1 | 0 | |
| NaiveRuntime (s)=0.5, Leak. (%)=0, # fails=02026.05 | 12.6 | -45.4 | |
| DriftRuntime (s)=0.5, Leak. (%)=0, # fails=02026.05 | 9.7 | -45.8 |