结合可视图的多状态交通流时间序列特性分析

发布时间：2018-06-25 04:44

本文选题：交通流时间序列 + 可视图　；参考：《物理学报》2017年23期

【摘要】：交通流时间序列的研究主要采用数据挖掘和机器学习的方法,这些"黑箱"挖掘方法很难直观反映序列特性.为增强交通流时间序列及其特征分析的可视化性,结合可视图理论来构建交通流时间序列的关联网络,从复杂网络角度实现交通流时间序列的特性分析.在网络构建的过程中,考虑到不同交通状态下交通流表征具有的差异性,首先利用交通流参量的相关性对交通流状态进行分类,然后构建不同交通状态下的时间序列复杂网络,并对这些网络的特征属性给出统计分析,如度分布、聚类系数、网络直径、模块化等.研究表明,可视图法可为交通流时间序列映射到网络提供有效途径,并且不同状态下交通流时间序列构建的复杂网络的模块化、聚类系数和度分布等统计特征呈现一定的变化规律,为交通流运行态势的研究提供了可视化的分析角度.
[Abstract]:The methods of data mining and machine learning are used in the study of traffic flow time series. These "black box" mining methods can hardly directly reflect the characteristics of traffic flow time series. In order to enhance the visualization of traffic flow time series and its characteristic analysis, the correlation network of traffic flow time series is constructed based on visual graph theory, and the characteristic analysis of traffic flow time series is realized from the point of view of complex network. In the process of network construction, considering the difference of traffic flow representation under different traffic conditions, the traffic flow state is classified by the correlation of traffic flow parameters, and then the complex network of time series under different traffic states is constructed. Statistical analysis of the characteristic attributes of these networks is given, such as degree distribution, clustering coefficient, network diameter, modularization and so on. The results show that the visual graph method can provide an effective way to map the traffic flow time series to the network and modularize the complex network constructed by the traffic flow time series under different states. The statistical features such as clustering coefficient and degree distribution show a certain variation law, which provides a visual analysis angle for the study of traffic flow situation.
【作者单位】：吉林大学交通学院;吉林化工学院信息与控制工程学院;吉林大学吉林省道路交通重点实验室;
【基金】：国家自然科学基金(批准号:51308248) 国家科技支撑计划(批准号:2014BAG03B03)资助的课题~~
【分类号】：O157.5;U491.112

【相似文献】