Traffic flow single-step forward forecasting on Highways England (test)
4.35RMSEHMDLF (with CNNGRU-Attention module)
Evaluation results
| Method | Links | |
|---|---|---|
| HMDLF (with CNNGRU-Attention module)Type=our model for multi-modality(flow/speed/journeytime), module=CNNGRU-Attention module2018.03 | 4.35 | |
| HMDLF (with CNNGRU module)Type=our model for multi-modality(flow/speed/journeytime), module=CNNGRU module2018.03 | 4.67 | |
| HMDLF (with CNNLSTM module)Type=our model for multi-modality(flow/speed/journeytime), module=CNNLSTM module2018.03 | 5.23 | |
| CNN-GRUType=baseline deep learning models for single modality (traffic flow)2018.03 | 9.09 | |
| CNNType=baseline deep learning models for single modality (traffic flow)2018.03 | 9.34 | |
| CNN-LSTMType=baseline deep learning models for single modality (traffic flow)2018.03 | 9.75 | |
| LSTMType=baseline deep learning models for single modality (traffic flow)2018.03 | 11.14 | |
| GRUType=baseline deep learning models for single modality (traffic flow)2018.03 | 11.15 | |
| RNNType=baseline deep learning models for single modality (traffic flow)2018.03 | 12.18 | |
| DTRType=shallow learning models for single modality (traffic flow)2018.03 | 12.51 | |
| LRType=shallow learning models for single modality (traffic flow)2018.03 | 12.54 | |
| RIDGEType=shallow learning models for single modality (traffic flow)2018.03 | 12.56 | |
| ARIMAType=shallow learning models for single modality (traffic flow)2018.03 | 14.01 | |
| SVR-RBFType=shallow learning models for single modality (traffic flow)2018.03 | 15.41 | |
| SVR-LINEARType=shallow learning models for single modality (traffic flow)2018.03 | 16.21 | |
| SVR-POLYType=shallow learning models for single modality (traffic flow)2018.03 | 41.64 |