|
| 摘要: |
| 本文研究了城市道路车辆的平均行驶速度在时间以及空间上的变化规律.利用道路交通的网络结构特征，提出了有向图上的高维时间序列模型.通过引入ℓ1--惩罚项，本文建立了正则化最小二乘的参数估计方法，并对此过程中调节参数的选择方式进行了研讨.进一步，本文将所提出的方法应用于成都市道路网络的城市链路速度的实例，并以成都市中心城区和偏远郊区为例，揭示了其交通运输效率在时间以及空间上的变化规律.该项研究可为疏导交通堵塞以及缓解交通压力提供有效的统计建议. |
| 关键词: 网络数据 车辆平均行驶速度 高维 交通 |
| DOI： |
| 分类号:O212 |
| 基金项目:福建省教育厅中青年基金（JAT200630）. |
|
| HIGH-DIMENSIONAL TIME SERIES MODEL ON DIRECTED GRAPH AND ITS APPLICATION IN TRAFFIC NETWORK |
|
XIANG Qiao-yi,QI Hui
|
| Abstract: |
| In this paper, we study the variation of the average speed of urban road vehicles in time and space. Based on the characteristics of road traffic network structure, the higher dimensions on the directed graph are proposed. By introducing ℓ1-penalty term, the regularized least squares parameter estimation method is established, and the selection method of adjusting parameters in this process is studied. The method is applied to the urban link speed data set of Chengdu road network. The temporal and spatial variation rules of transportation efficiency in downtown and remote suburbs of Chengdu are obtained. This study can provide effective statistical suggestions for relieving traffic jams and alleviating traffic pressure. |
| Key words: network data average vehicle speed high dimensions transportation |
