基于数据挖掘技术的交通流数据分析研究  

Analysis of Traffic Flow Data Based on Data Mining

作　　者：许兆霞[1] 王磊[1] 林勇 李树彬[1,2] 张晓波[1] 

机构地区：[1]山东省科学院自动化研究所,山东济南250014 [2]北京交通大学交通运输学院,北京100044

出　　处：《淮海工学院学报（自然科学版）》2009年第1期31-34,共4页Journal of Huaihai Institute of Technology:Natural Sciences Edition

基　　金：公安部应用创新计划项目(2007YYCXSDST057)

摘　　要：利用数据挖掘技术,通过对历史数据的分析预测下一个时间间隔的交通流状况,可以为交通流诱导和信息发布打下基础;通过对路口流量历史数据的聚类分析可得出单路口TOD控制算法的最优时段分段和各时段中的最优控制参数,从而优化单路口控制算法的控制效果;通过对路段流量历史数据之间的关联分析,可得出路段之间的关联规则,从而可以由一个路段的流量推断其关联路段的流量,为实时交通流诱导和信息发布提供实时依据。The prediction of traffic flow at the next time interval based on the analysis of historic data and the application of data mining technology is of great help in traffic division and information distribution. The optimal time distribution by TOD control algorithm at single junctions and the optimal control parameters can be obtained by cluster analysis of historical data at intersections, and in this case, the control algorithm can be optimized. Correlation rules can be drawn by correlation analysis of historical data for traffic at different road sections, and hence the traffic of correlated road sections can be estimated by specific traffic at a known section. Such an estimation can provide references for the real-time traffic division and information distribution.

关 键 词：数据挖掘 交通流 预测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]