基于聚类分析的营运驾驶人危险驾驶行为研究  

Study on Risky Behavior of Commercial Drivers Based on Clustering Analysis

作　　者：牛增良[1] 李海斌[2] 王文峰[2] 白丽霞[3] 

机构地区：[1]长安大学汽车运输保障技术交通行业重点实验室,陕西西安710064 [2]中国汽车技术研究中心北京中机车辆司法鉴定所,北京100176 [3]上海市政工程设计研究总院(集团)有限公司天津分公司,天津300042

出　　处：《山东交通学院学报》2014年第1期19-23,共5页Journal of Shandong Jiaotong University

摘　　要：为改善交通安全状况,减少营运客车交通事故,从危险驾驶行为角度出发研究外部影响因素与驾驶行为之间的关系。以大量重特大交通事故数据为基础,定量分析聚类方法在危险驾驶行为研究中的应用。运用模糊聚类、系统聚类2种方法对危险驾驶行为划分层次进行分析,并通过营运客车驾驶人危险驾驶行为的实例对2种方法得出的结果进行讨论。研究表明,通过聚类方法归并为一类的危险驾驶行为,与对应的外部影响因素有相近的关联度。In order to improve the traffic safety and decrease the amount of traffic accidents about commercial vehicles,this article begins with studying the relationship between the external factors and driving behaviors from the risky driving behaviors.On the basis of a large number of traffic accident data,this paper analyzes the application of the quantificationally clustering method to the risky driving behavior research.The article puts forward the thought of applying the hierarchical clustering and fussy clustering to the calcification analysis of the risky driving behaviors.Also the paper discusses the result obtained from the two methods through the practical example of the risky driving behaviors of commercial drivers.The research shows that the merged risky driving behavior by the clustering method has the similar relevancy to the responding external factors.

关 键 词：道路安全 交通事故 营运客车 驾驶行为 聚类分析 

分 类 号：U471.3[机械工程—车辆工程]