Spatial-temporal Time Series Forecasting on PEMS08
15.72MAESCINet
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SCINet2021.06 | 15.72 | 9.8 | 24.76 | 1.44 | |
| AGCRN2021.06 | 15.95 | 10.09 | 25.22 | — | |
| STFGNN2021.06 | 16.64 | 10.6 | 26.22 | — | |
| STSGCN2021.06 | 17.13 | 10.96 | 26.8 | — | |
| LSGCN2021.06 | 17.73 | 11.2 | 26.76 | — | |
| DCRNN2021.06 | 17.86 | 11.45 | 27.83 | — | |
| STGCN2021.06 | 18.02 | 11.4 | 27.83 | — | |
| ASTGCN(r)2021.06 | 18.61 | 13.08 | 28.16 | — | |
| Graph WaveNet2021.06 | 19.13 | 12.68 | 31.05 | — | |
| STDNTraining Loss=FreST Loss2026.02 | 20.502 | — | 33.393 | — | |
| TCNre-implemented=true, variant=normal convolutions2021.06 | 21.42 | 13.09 | 34.03 | — | |
| STDNTraining Loss=MSE Loss2026.02 | 21.434 | — | 33.867 | — | |
| LSTMre-implemented=true2021.06 | 22.2 | 15.32 | 32.06 | — | |
| TCNre-implemented=true2021.06 | 22.72 | 14.03 | 35.79 | — | |
| StaeformerTraining Loss=FreST Loss2026.02 | 29.551 | — | 41.81 | — | |
| STGCNTraining Loss=FreST Loss2026.02 | 32.072 | — | 49.615 | — | |
| StaeformerTraining Loss=MSE Loss2026.02 | 33.373 | — | 47.504 | — | |
| STGCNTraining Loss=MSE Loss2026.02 | 34.219 | — | 54.567 | — | |
| StemGNNTraining Loss=MSE Loss2026.02 | 40.097 | — | 78.742 | — | |
| StemGNNTraining Loss=FreST Loss2026.02 | 41.378 | — | 60.38 | — | |
| STIDTraining Loss=MSE Loss2026.02 | 42.573 | — | 62.418 | — | |
| STIDTraining Loss=FreST Loss2026.02 | 43.904 | — | 62.209 | — |