摘要
利用数据挖掘技术从国家安全生产监督管理总局事故查询系统中我国2004—2014年间发生的重大交通事故中筛选出具有道路类型信息的2 383件交通事故数据,按照省级行政区和道路类型进行聚类分析,将我国31个省级行政区域的道路交通安全水平进行了归类,并利用相关性分析法分析了社会经济发展水平、地势阶梯和人口密度等区域特征与道路交通安全水平之间的关系。结果表明:通过文本数据挖掘技术可以得到对我国道路交通安全评价有价值的信息;我国68%的省级行政区域在不同道路类型上的道路交通安全水平存在差异性;利用相关性分析发现我国道路交通安全水平与社会经济发展水平、地势阶梯、人口密度等区域特征有较强的相关性。
This paper uses data mining technology to extract information from the Accident Query System operated by State Administration of Work Safety.The paper selects 2383 sever traffic accidents between 2004 and 2014 with the information of road type.Then the paper uses cluster analysis to get the cluster center for classifying the traffic safety levels,and applies the levels to 31 provincial regions according to the road types.Also,the paper uses correlation analysis to find the relations between traffic safety level and regional characteristics,such as economical developing level,terrain ladder,and population density.The results show that much valuable information which can be helpful in traffic accident prevention is obtained by using the data mining technology.According to the classification of traffic safety level,it is found that the safety levels vary with road type in 68%of the provincial regions.And the correlation analysis shows strong correlations between road traffic safety levels and the regional characteristics.
作者
张坤
梅诗冬
景国勋
三上喜贵
ZHANG Kun;MEI Shidong;JING Guoxun;MIKAMI Yoshiki(School of Safety Science and Engineering,Henan Polytechnic University,Jiaozuo 454000, China;Department of Nuclear System Safety,Nagaoka University of Technology,Nagaoka 940-2188,Japan;Anyang Institute of Technology,Anyang 455000,China)
出处
《安全与环境工程》
CAS
北大核心
2018年第1期76-81,共6页
Safety and Environmental Engineering
基金
国家自然科学基金项目(51474098)
日本学术振兴会(JSPS)特别研究员奖励费基金项目(JP26-04306)
河南理工大学博士基金项目(B2013-008)
关键词
数据挖掘
道路交通
安全评价
聚类分析
相关性分析
data mining
road traffic
safety assessment
cluster analysis
correlation analysis
