| Autor: |
Zhang, Gongquan, Jin, Jieling, Chang, Fangrong, Huang, Helai |
| Zdroj: |
International Journal of Transportation Science and Technology; 20240101, Issue: Preprints |
| Abstrakt: |
•Uses real-time video data and deep learning for traffic conflict prediction.•Proposes a deep and cross network (DCN) model with lane-level traffic parameters.•Uses SHAP to explain the impact of dynamic traffic parameters on conflicts.•DCN model outperforms statistical and machine learning models in real-time prediction. |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
