Research Article
Road Travel Time Prediction Based on Improved Graph Convolutional Network
Table 1
Performance comparison of prediction models.
| Model | 15 min | 30 min | MAE | MAPE (%) | RMSE | MAE | MAPE (%) | RMSE |
| HA | 0.0401 | 25.37 | 0.0689 | 0.0431 | 28.48 | 0.0703 | ARIMA | 0.0394 | 23.96 | 0.0663 | 0.0419 | 25.06 | 0.0691 | LSTM | 0.0365 | 20.04 | 0.0624 | 0.0400 | 22.97 | 0.0652 | IGC-Net_T | 0.0326 | 18.57 | 0.0643 | 0.0387 | 21.18 | 0.0640 | STGCN | 0.0327 | 18.41 | 0.0638 | 0.0375 | 20.03 | 0.0653 | DCRNN | 0.0301 | 16.78 | 0.0611 | 0.0336 | 18.37 | 0.0641 | Graph WaveNet | 0.0293 | 16.21 | 0.0602 | 0.0317 | 17.11 | 0.0634 | IGC-Net_E | 0.0287 | 14.83 | 0.0596 | 0.0298 | 15.97 | 0.0610 | IGC-Net | 0.0269 | 13.70 | 0.0527 | 0.0292 | 15.49 | 0.0601 |
|
|