Time Series Forecasting on Traffic (test)
0.0592MSECo-TSFA
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Co-TSFAScenario Category=Continuous, Train Contamination=none, Test Contamination=input-only, Anomaly Type=-, Ratio (%)=-2025.12 | 0.0592 | — | 0.1633 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=const, Ratio (%)=102025.12 | 0.0629 | — | 0.1675 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=const, Ratio (%)=302025.12 | 0.07 | — | 0.1805 | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=none, Test Contamination=input+output, Anomaly Type=-, Ratio (%)=-2025.12 | 0.0722 | — | 0.1824 | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=continuous, Test Contamination=input-only, Anomaly Type=-, Ratio (%)=-2025.12 | 0.0731 | — | 0.1862 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input-only, Anomaly Type=const, Ratio (%)=102025.12 | 0.0748 | — | 0.1875 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input-only, Anomaly Type=const, Ratio (%)=302025.12 | 0.0795 | — | 0.1968 | — | — | — | |
| D³VAEPrediction Length (l_y)=82023.01 | 0.081 | 0.207 | — | — | — | — | |
| D³VAEPrediction Length (l_y)=162023.01 | 0.081 | 0.2 | — | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input+output, Anomaly Type=const, Ratio (%)=102025.12 | 0.0825 | — | 0.1996 | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=continuous, Test Contamination=input+output, Anomaly Type=-, Ratio (%)=-2025.12 | 0.0876 | — | 0.2064 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=missing, Ratio (%)=102025.12 | 0.0882 | — | 0.2093 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input+output, Anomaly Type=const, Ratio (%)=302025.12 | 0.0946 | — | 0.2204 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input-only, Anomaly Type=missing, Ratio (%)=102025.12 | 0.0966 | — | 0.2241 | — | — | — | |
| TimePrisminput length=nominal forecast horizon T2025.09 | 0.0983 | — | — | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input-only, Anomaly Type=const, Ratio (%)=102025.12 | 0.0997 | — | 0.2247 | — | — | — | |
| RobustTSFScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=missing, Ratio (%)=102025.12 | 0.1078 | — | 0.193 | — | — | — | |
| RobustTSFScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=Gaussian, Ratio (%)=302025.12 | 0.1095 | — | 0.1959 | — | — | — | |
| RobustTSFScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=Gaussian, Ratio (%)=102025.12 | 0.1096 | — | 0.1904 | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input+output, Anomaly Type=const, Ratio (%)=102025.12 | 0.1097 | — | 0.2385 | — | — | — | |
| RobustTSFScenario Category=Continuous, Train Contamination=none, Test Contamination=input-only, Anomaly Type=-, Ratio (%)=-2025.12 | 0.1101 | — | 0.2029 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=Gaussian, Ratio (%)=102025.12 | 0.1117 | — | 0.2468 | — | — | — | |
| RobustTSFScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=const, Ratio (%)=302025.12 | 0.1126 | — | 0.2029 | — | — | — | |
| RobustTSFScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=const, Ratio (%)=102025.12 | 0.1133 | — | 0.1982 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input-only, Anomaly Type=Gaussian, Ratio (%)=102025.12 | 0.1179 | — | 0.2548 | — | — | — | |
| RobustTSFScenario Category=Continuous, Train Contamination=none, Test Contamination=input+output, Anomaly Type=-, Ratio (%)=-2025.12 | 0.1184 | — | 0.2152 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input+output, Anomaly Type=missing, Ratio (%)=102025.12 | 0.1325 | — | 0.2701 | — | — | — | |
| RobustTSFScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=missing, Ratio (%)=302025.12 | 0.1401 | — | 0.2346 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input-only, Anomaly Type=missing, Ratio (%)=302025.12 | 0.143 | — | 0.2903 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=missing, Ratio (%)=302025.12 | 0.149 | — | 0.2982 | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input-only, Anomaly Type=const, Ratio (%)=302025.12 | 0.158 | — | 0.3065 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input+output, Anomaly Type=Gaussian, Ratio (%)=102025.12 | 0.1611 | — | 0.3046 | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input+output, Anomaly Type=const, Ratio (%)=302025.12 | 0.1629 | — | 0.3107 | — | — | — | |
| RobustTSFScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input-only, Anomaly Type=const, Ratio (%)=102025.12 | 0.1679 | — | 0.2748 | — | — | — | |
| TimeGrad2025.09 | 0.19 | — | — | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input-only, Anomaly Type=Gaussian, Ratio (%)=302025.12 | 0.1942 | — | 0.3483 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=none, Test Contamination=-, Anomaly Type=Gaussian, Ratio (%)=302025.12 | 0.1956 | — | 0.3544 | — | — | — | |
| RobustTSFScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input+output, Anomaly Type=const, Ratio (%)=102025.12 | 0.1985 | — | 0.3007 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input+output, Anomaly Type=missing, Ratio (%)=302025.12 | 0.2028 | — | 0.3433 | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input-only, Anomaly Type=missing, Ratio (%)=102025.12 | 0.2123 | — | 0.3124 | — | — | — | |
| Co-TSFAScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input+output, Anomaly Type=missing, Ratio (%)=102025.12 | 0.2153 | — | 0.3196 | — | — | — | |
| RobustTSFScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input-only, Anomaly Type=const, Ratio (%)=302025.12 | 0.2234 | — | 0.3467 | — | — | — | |
| Co-TSFAScenario Category=Pointwise, Train Contamination=continuous, Test Contamination=input+output, Anomaly Type=Gaussian, Ratio (%)=302025.12 | 0.2649 | — | 0.4011 | — | — | — | |
| TBP m*N=192, burn-in length (m)=20, stateful=false2026.02 | 0.276 | — | — | — | -3 | — | |
| TBP m*N=96, burn-in length (m)=75, stateful=false2026.02 | 0.28 | — | — | — | -8 | — | |
| RobustTSFScenario Category=Continuous, Train Contamination=pointwise, Test Contamination=input+output, Anomaly Type=const, Ratio (%)=302025.12 | 0.2811 | — | 0.3894 | — | — | — | |
| TBP m-barN=192, burn-in length (m)=191, stateful=false2026.02 | 0.284 | — | — | — | — | — | |
| TBP m-barN=96, burn-in length (m)=95, stateful=false2026.02 | 0.305 | — | — | — | — | — | |
| TBP m*N=48, burn-in length (m)=0, stateful=false2026.02 | 0.316 | — | — | — | -8 | — | |
| TimeMCLinput length=historical context longer than T2025.09 | 0.319 | — | — | — | — | — | |
| TBP m-barN=48, burn-in length (m)=47, stateful=false2026.02 | 0.343 | — | — | — | — | — | |
| SFFData proportion=75%, Prediction length=962026.06 | 0.3478 | — | — | — | — | — | |
| SFFData proportion=25%, Prediction length=962026.06 | 0.3488 | — | — | — | — | — | |
| SFFData proportion=50%, Prediction length=962026.06 | 0.3497 | — | — | — | — | — | |
| SFFData proportion=100%, Prediction length=962026.06 | 0.3551 | — | — | — | — | — | |
| TFSData proportion=50%, Prediction length=962026.06 | 0.3552 | — | — | — | — | — | |
| FFData proportion=25%, Prediction length=962026.06 | 0.3582 | — | — | — | — | — | |
| FFData proportion=50%, Prediction length=962026.06 | 0.3586 | — | — | — | — | — | |
| TBP sfN=192, burn-in length (m)=191, stateful=true2026.02 | 0.359 | — | — | — | — | — | |
| FFData proportion=100%, Prediction length=962026.06 | 0.3599 | — | — | — | — | — | |
| TFSData proportion=75%, Prediction length=962026.06 | 0.3606 | — | — | — | — | — | |
| TFSData proportion=100%, Prediction length=962026.06 | 0.3609 | — | — | — | — | — | |
| FFData proportion=75%, Prediction length=962026.06 | 0.361 | — | — | — | — | — | |
| TSLANetH=96, L=7202025.09 | 0.362 | — | — | — | — | — | |
| PatchTSTPrediction Length=962026.06 | 0.364 | — | — | — | — | — | |
| PatchTSTH=96, L=7202025.09 | 0.366 | — | — | — | — | — | |
| PatchTST (Sup.)Prediction Horizon=96, Evaluation Protocol=Supervised2022.11 | 0.367 | — | 0.251 | — | — | — | |
| Tactis22025.09 | 0.368 | — | — | — | — | — | |
| TFSData proportion=25%, Prediction length=962026.06 | 0.3688 | — | — | — | — | — | |
| Timer-XLPrediction Length (T)=Avg., Look-back Length (L)=Best2026.01 | 0.374 | — | 0.255 | — | — | — | |
| TSLANetH=192, L=7202025.09 | 0.377 | — | — | — | — | — | |
| TimeAlignConfidence Interval=99%2025.09 | 0.378 | — | 0.24 | — | — | — | |
| PatchTSTPrediction Length=1922026.06 | 0.379 | — | — | — | — | — | |
| TimesNetPrediction Length=962026.06 | 0.381 | — | — | — | — | — | |
| PatchTST (Sup.)Prediction Horizon=192, Evaluation Protocol=Supervised2022.11 | 0.385 | — | 0.259 | — | — | — | |
| RRRH=96, L=7202025.09 | 0.385 | — | — | — | — | — | |
| FITSH=96, L=7202025.09 | 0.386 | — | — | — | — | — | |
| Root PurgeH=96, L=7202025.09 | 0.386 | — | — | — | — | — | |
| PatchTST (Fine-tuning)Prediction Horizon=96, Pre-training Dataset=Electricity, Evaluation Protocol=Fine-tuning2022.11 | 0.388 | — | 0.273 | — | — | — | |
| PatchTSTH=192, L=7202025.09 | 0.388 | — | — | — | — | — | |
| SparseTSFH=96, L=7202025.09 | 0.389 | — | — | — | — | — | |
| PatchTST/64Prediction Length (T)=Avg., Look-back Length (L)=3362026.01 | 0.391 | — | 0.264 | — | — | — | |
| TSLANetH=336, L=7202025.09 | 0.391 | — | — | — | — | — | |
| FilterNetH=96, L=7202025.09 | 0.393 | — | — | — | — | — | |
| RRRH=192, L=7202025.09 | 0.396 | — | — | — | — | — | |
| TVNetConfidence Interval=99%2025.09 | 0.396 | — | 0.268 | — | — | — | |
| PatchTSTPrediction Length=3362026.06 | 0.396 | — | — | — | — | — | |
| TimesNetPrediction Length=1922026.06 | 0.396 | — | — | — | — | — | |
| FITSH=192, L=7202025.09 | 0.397 | — | — | — | — | — | |
| Root PurgeH=192, L=7202025.09 | 0.397 | — | — | — | — | — | |
| PatchTST (Sup.)Prediction Horizon=336, Evaluation Protocol=Supervised2022.11 | 0.398 | — | 0.265 | — | — | — | |
| N-BEATSHorizon=962023.07 | 0.398 | — | — | — | — | — | |
| SparseTSFH=192, L=7202025.09 | 0.398 | — | — | — | — | — | |
| PatchTSTH=336, L=7202025.09 | 0.398 | — | — | — | — | — | |
| DLinearPrediction Length=962026.06 | 0.399 | — | — | — | — | — | |
| PatchTST (Fine-tuning)Prediction Horizon=192, Pre-training Dataset=Electricity, Evaluation Protocol=Fine-tuning2022.11 | 0.4 | — | 0.277 | — | — | — | |
| PatchTST (Lin. Prob.)Prediction Horizon=96, Pre-training Dataset=Electricity, Evaluation Protocol=Linear Probing2022.11 | 0.4 | — | 0.288 | — | — | — | |
| TimesNetPrediction Length=Avg2026.06 | 0.4 | — | — | — | — | — | |
| FITSHorizon=962023.07 | 0.401 | — | — | — | — | — | |
| N-HiTSHorizon=962023.07 | 0.402 | — | — | — | — | — |