山东科学 ›› 2024, Vol. 37 ›› Issue (1): 107-117.doi: 10.3976/j.issn.1002-4026.20230058

• 交通运输 • 上一篇    下一篇

基于复杂网络的交通序列数据特性

孟勃1(), 孔祥科2, 李树彬3,*()   

  1. 1.中国人民公安大学 犯罪学学院,北京 100038
    2.山东轨道交通勘察设计院有限公司,山东 济南 250014
    3.山东警察学院 道路交通安全研究所,山东 济南 250014
  • 收稿日期:2023-04-09 出版日期:2024-02-20 发布日期:2024-01-26
  • 通信作者: 李树彬 E-mail:bo13963591527@163.com;li_shu_bin@163.com
  • 作者简介:孟勃(1986—),女,博士研究生,研究方向为交通犯罪学。E-mail: bo13963591527@163.com
  • 基金资助:
    国家自然科学基金(71871130);国家自然科学基金(719711254);山东省公安厅科技专项(SDGP370000000202302006727A001);山东公安科技创新计划项目(GAKJCX2022-1)

The characteristics of traffic sequence data based on complex network

MENG Bo1(), KONG Xiangke2, LI Shubin3,*()   

  1. 1. College of Criminology, People's Public Security University of China, Beijing 100038,China
    2. Shandong Rail Transit Survey and Design Institute Co., Ltd., Jinan 250014, China
    3. Institute of Road Traffic Safety, Shandong Police College, Jinan 250014, China
  • Received:2023-04-09 Online:2024-02-20 Published:2024-01-26
  • Contact: LI Shubin E-mail:bo13963591527@163.com;li_shu_bin@163.com

摘要:

为了进一步研究交通流特性,采用复杂网络方法对交通序列数据进行分析。提出了箱型图-聚类算法模型用于识别和填充初始数据中的缺失值和异常值;通过相空间重构方法将一维数据重构为网络节点,选取连接阈值确定网络节点的连接关系,将交通序列数据构建为复杂网络,对复杂网络的结构和定量指标进行分析。研究结果表明交通序列数据复杂网络的结构一定程度上可以反映路段的交通流状态。该结果有助于优化数据预处理方法,拓展复杂网络在交通序列数据研究中的应用。

关键词: 复杂网络, 数据分析, 网络构建方法, 相空间重构, 聚类算法

Abstract:

To study the traffic flow characteristics, the traffic data is analyzed using a complex network method. A box plot-clustering algorithm model is proposed to identify and fill in missing values and outliers in the initial data. The one-dimensional data is reconstructed into network nodes using the phase space reconstruction method. Additionally, the connection threshold is selected to determine the connection relationship of network nodes to convert the traffic sequence data as a complex network and analyze the structure and quantitative indicators of the network. The result shows that the structure of the complex network of traffic data can reflect the traffic flow state of the road section to a certain extent. The research optimizes the data preprocessing method and extends the application of complex networks into traffic data research.

Key words: complex networks, data analysis, network building methods, phase space reconstruction, clustering algorithm

中图分类号: 

  • U491