Probabilistic Forecasting on wiki
0.214CRPSDeepState
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DeepStatemodel=state space model2023.07 | 0.214 | — | |
| TSDiff-Condvariant=conditional2023.07 | 0.218 | — | |
| MQ-CNNarchitecture=CNN-based2023.07 | 0.22 | — | |
| TSDiff-Qguidance=quantile guidance2023.07 | 0.221 | — | |
| TFTarchitecture=Temporal Fusion Transformer2023.07 | 0.229 | — | |
| DeepARarchitecture=RNN-based2023.07 | 0.231 | — | |
| Transformerarchitecture=self-attention-based2023.07 | 0.231 | — | |
| TSDiff-MSguidance=mean square guidance2023.07 | 0.257 | — | |
| CSDImodel=conditional diffusion2023.07 | 0.289 | — | |
| Seasonal Naive2023.07 | 0.41 | — | |
| TACTiS-22025.09 | 0.484 | — | |
| TimePrismNumber of scenarios (N)=625, Input length=302025.09 | 0.506 | — | |
| TimeGrad2025.09 | 0.517 | — | |
| TimeMCLNumber of hypotheses=162025.09 | 0.64 | — | |
| TimePrism-16Number of scenarios (N)=16, Input length=302025.09 | 0.654 | — | |
| ETS2023.07 | 0.715 | — | |
| TempFlowBackbone=LSTM2025.09 | 1.26 | — | |
| Linear2023.07 | 1.624 | — | |
| Trf.FlowBackbone=Transformer2025.09 | 1.71 | — | |
| DeepAR2025.09 | 1.75 | — | |
| ETS2025.09 | 4.88 | — | |
| TimePreNumber of hypotheses=162025.11 | 6.14 | — | |
| Tactis2Number of hypotheses=162025.11 | 6.24 | — | |
| TempFlowNumber of hypotheses=162025.11 | 14.49 | — | |
| TimeMCL (R.)Number of hypotheses=162025.11 | 14.5 | — | |
| DeepARHypotheses=162025.11 | — | 382,340 | |
| ETSHypotheses=162025.11 | — | 835,095 | |
| Tactis2Hypotheses=162025.11 | — | 263,975 | |
| TempFlowHypotheses=16, Backbone=LSTM2025.11 | — | 395,996 | |
| TempFlow (Trf.)Hypotheses=16, Backbone=Transformer2025.11 | — | 561,226 | |
| TimeGradHypotheses=162025.11 | — | 275,437 | |
| TimeMCL (A.)Hypotheses=16, Training=Annealed-MCL2025.11 | — | 276,315 | |
| TimeMCL (R.)Hypotheses=16, Training=Relaxed-WTA2025.11 | — | 268,832 | |
| TimePreHypotheses=162025.11 | — | 263,492 |