Demand Forecasting on SNBC industrial dataset
1.326MSEValidation-selected Oracle Reference
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Validation-selected Oracle ReferenceExec.=1.02026.06 | 1.326 | 1.097 | 12.8 | 16 | |
| FAME Top-2Exec.=1.92, Norm. Cost=2.42026.06 | 1.384 | 1.143 | 13.3 | 12.4 | |
| FAME-CostAwareExec.=1.54, Norm. Cost=1.82026.06 | 1.397 | 1.155 | 13.5 | 11.5 | |
| FAME Top-1Exec.=1.0, Norm. Cost=1.32026.06 | 1.421 | 1.174 | 13.7 | 10 | |
| Dense Soft MoEExec.=10.0, Norm. Cost=16.52026.06 | 1.438 | 1.191 | 13.9 | 8.9 | |
| Stacking EnsembleExec.=10.0, Norm. Cost=17.02026.06 | 1.472 | 1.218 | 14.2 | 6.8 | |
| Rule-based USFFExec.=1.0, Norm. Cost=1.22026.06 | 1.489 | 1.239 | 14.4 | 5.7 | |
| FFORMA-style WeightingExec.=10.0, Norm. Cost=16.42026.06 | 1.514 | 1.255 | 14.6 | 4.1 | |
| Cluster-then-ForecastExec.=1.0, Norm. Cost=1.22026.06 | 1.523 | 1.274 | 14.8 | 3.6 | |
| Uniform EnsembleExec.=10.0, Norm. Cost=16.42026.06 | 1.566 | 1.302 | 15.1 | 0.8 | |
| LightGBMExec.=1.0, Norm. Cost=1.02026.06 | 1.579 | 1.348 | 15.7 | 0 | |
| XGBoostExec.=1.0, Norm. Cost=1.12026.06 | 1.714 | 1.438 | 16.8 | -8.5 | |
| Linear RegressionExec.=1.0, Norm. Cost=0.62026.06 | 1.798 | 1.486 | 17.3 | -13.9 | |
| TimesNetExec.=1.0, Norm. Cost=4.12026.06 | 1.833 | 1.61 | 17.9 | -16.1 | |
| TimeMixerExec.=1.0, Norm. Cost=3.62026.06 | 1.845 | 1.635 | 18.1 | -16.8 | |
| Holt-WintersExec.=1.0, Norm. Cost=0.82026.06 | 4.307 | 3.624 | 38.4 | -172.7 | |
| SARIMAExec.=1.0, Norm. Cost=0.82026.06 | 4.798 | 4.044 | 41.8 | -203.8 |