| Model Name | Type | Data Type | 30days-10mins | 60days-10mins | 60days-1hour |
|---|---|---|---|---|---|
| MAE| RMSE | MAE| RMSE | MAE| RMSE | |||
| HI | Time Series | Time Series | 30.606| - | 12.560| 17.628 | 90.631| 128.092 |
| ARIMA | Time Series | Time Series | 28.024| - | -| - | -| - |
| Theta | Time Series | Time Series | 27.230| - | -| - | -| - |
| LSTM | Time Series + | Sliding Window | 39.780| 88.138 | -| - | 227.606| 526.731 |
| CNN-LSTM | Spatio-temporal | Sliding Window | -| - | -| - | -| - |
| STN | Spatio-temporal | Sliding Window | 25.843| 50.051 | 12.014| 22.443 | 96.799| 173.383 |
| Vanilla TransformerE | Time Series + | Sliding Window | 28.745| 54.207 | -| - | -| - |
| Informer | Time Series + | Sliding Window | 26.003| 53.063 | 11.568 | 24.009 | 74.6274 | 171.131 |
| StTran | Spatio-temporal | Sliding Window | 27.875| 56.149 | 11.896| 24.597 | -| - |
| ViT | Spatio-temporal | Sliding Window | 25.414 | 53.514 | -| - | -| - |
| - | - | - | - | - | - |
| MLP | Time Series | Full grid | 26.562| 52.882 | -| - | -| - |
| ST-DenseNet | Spatio-temporal | Full grid | 46.952| 154.664 | 33.475| 75.603 | 203.403| 517.890 |
| MVSTGN | Spatio-temporal | Full grid | 47.848| 147.858 | 16.973| 56.853 | 156.780| 408.457 |
| STGCN | Spatio-temporal | Full grid | 28.662| 58.667 | 14.742| 40.747 | 255.056| 480.541 |
| ViT-Pyramid | Spatio-temporal | Full grid | 27.115| 53.869 | 24.239| 41.66 | |
| ViT | Spatio-temporal | Full grid | 25.84|51.563 | 11.649| 20.96 | 71.639| 164.173 |
| Subject | Name | Temporal Resolution | Statistics | Description | Type |
|---|---|---|---|---|---|
| Traffic | METR-LA | 5 minutes | (34,272, 207) | traffic speed data | MTS |
| Traffic | PEMS-BAY | 5 minutes | (52,116, 326) | Traffic speed and flow | MTS |
| Traffic | PEMS-2 | 5 minutes | (61,056, 325) | Traffic speeds | MTS |
| Traffic | LargeST | 5 minutes | (525,888, 8600) | Traffic sensors | MTS |
| Traffic | TrafficBJ | 5 minutes | (21,600, 3126) | Traffic speeds | MTS |
| Traffic | TaxiBJ | 30 minutes | 4*(7220, 2, 32, 32) | Crowd flows | Tensor |
| Traffic | JONAS-NYC | 30 minutes | 2*(4800, 16, 8, 2) | Demand & Supply | Tensor |
| Traffic | JONAS-DC | 1 hour | 2*(2400, 9, 12, 2) | Demand & Supply | Tensor |
| Traffic | COVID-CHI | 2 hours | 3*(6600, 14, 8, 2) | Demand & Supply | Tensor |
| Traffic | COVID-US | 1 hour | (4800, 51, 10) | Travel purpose | Tensor |
| Traffic | BikeNYC | -| - | trip records from June,2013 to January,2024 | Bike trip records | Record |
| Traffic | TaxiNYC | -| - | 22,394,490 trip records | Taxi trip records | Record |
| Finance | M4 | 1 hour ~ 1 year | 100,000 time series | Time series | TS |
| Finance | M5 | 1 day | (30490, 1947), (30490, 1919) (train & valid) | Walmart sales forecast | MTS |
| Finance | NASDAQ 100 | 1 minute | (391191, 104), (390191, 104) | Stock prices | MTS |
| Finance | Stock Market Dataset | 1 day | 8049*(D, 7) | NASDAQ stock prices | MTS |
| Finance | CrypTop12 | 1 day | (1255, 12, 7)& Tweets | NASDAQ stock prices | Tensor |
| Finance | stocknet-dataset | 1 day | (88, 731, 5) | Stock price movement | Tensor |
| Finance | CSI300 | 1 day | Stock market index | TS | |
| Telecom | Milan&Trentino | 15 minutes | (T, 100, 100, 5) | 5 types of CDRs | Tensor |
| Health | PTB-XL | 2 ms, 10 ms | Sample rate 100Hz, 500Hz for 10s | ECG dataset | TS |
| Health | MIMIC-III | various | Clinical data | Series | |
| Health | MIT-BIH Arrhythmia Database | 2.78 ms | 48 records, 2 input channels | ECG dataset | MTS |
| Health | PTB Diagnostic ECG Database | 1 ms | 549 records, 16 input channels | ECG dataset | MTS |
| Weather | Shifts-Weather | -| - | (3129592, 129) | Weather prediction | MTS |
| Energy | ElectricityLoad | 15 minutes | (140256,370) | Electricity consumption | MTS |