Covariate Selection for Forecasting on M5
0.92Mean NDCGOurs-Cluster
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Ours-ClusterModel=FMTimeSeries2026.06 | 0.92 | 47 | |
| OursModel=FMTimeSeries2026.06 | 0.89 | 8 | |
| RossModel=FMTimeSeries2026.06 | 0.88 | 27 | |
| Ours-ClusterModel=CatBoostRegressor2026.06 | 0.86 | 63 | |
| Ours-ClusterModel=Chronos-22026.06 | 0.83 | 73 | |
| Ours-ClusterModel=DeepAR2026.06 | 0.82 | 76 | |
| OursModel=CatBoostRegressor2026.06 | 0.81 | 15 | |
| OursModel=Chronos-22026.06 | 0.78 | 11 | |
| OursModel=DeepAR2026.06 | 0.76 | 3 | |
| RossModel=CatBoostRegressor2026.06 | 0.75 | 17 | |
| MixtureModel=FMTimeSeries2026.06 | 0.74 | 18 | |
| RossModel=DeepAR2026.06 | 0.71 | 17 | |
| RossModel=Chronos-22026.06 | 0.7 | 9 | |
| MixtureModel=CatBoostRegressor2026.06 | 0.62 | 5 | |
| MixtureModel=Chronos-22026.06 | 0.59 | 7 | |
| MixtureModel=DeepAR2026.06 | 0.54 | 4 |