Table 2.
Verification results of LTCOP-LSTM model in different scenarios.
| Dataset | Prediction time window (min) | MAE (vehicles) | RMSE (vehicles) | Training set size (samples) | Testing set size (samples) | Prediction frequency (s) | Root mean square relative error (%) |
|---|---|---|---|---|---|---|---|
| METR-LA | 10 | 5.613 | 7.844 | 3000 | 1200 | 30 | 3.87 |
| PeMS | 15 | 4.912 | 6.29 | 2500 | 1000 | 60 | 2.76 |
| Taxi Trajectory | 20 | 6.721 | 9.301 | 2800 | 1100 | 45 | 4.12 |
| ULTRA | 25 | 5.008 | 7.617 | 3500 | 1400 | 30 | 3.25 |
| METR-LA | 30 | 7.234 | 9.835 | 3200 | 1300 | 180 | 5.47 |
| PeMS | 10 | 5.000 | 6.893 | 3100 | 1200 | 45 | 3.00 |
| Taxi Trajectory | 35 | 8.146 | 10.568 | 2900 | 1150 | 60 | 4.78 |
| ULTRA | 15 | 4.897 | 6.151 | 3300 | 1250 | 30 | 2.65 |