基于交通安全数据挖掘的高速公路事故多发点研究

发布时间：2018-04-27 23:37

本文选题：交通安全 + 高速公路　；参考：《山东科技大学》2017年硕士论文

【摘要】：我国的十三五规划纲要对交通行业提出了新的要求:到2020年,基本建成安全、便捷、高效、绿色的现代综合交通运输体系。目前我国高速公路通车总里程数居世界第一位,机动车保有量也大幅增加,交通安全问题日益凸显。道路交通安全已经成为广受社会各界所关注的社会焦点问题。论文在对山东某高速公路事故发生量进行统计分析的基础上,总结出高速公路交通事故特性,主要从交通安全事故的时间特性、空间特性、气候分布特性和形态特性等方面进行展开研究,并在此基础上分析得出高速公路交通安全影响因素。论文对区域高速公路路网交通安全数据的多源异构性进行分析,运用数据融合的方法进行预处理,并在此基础上研究数据挖掘获取信息的过程,建立统一的数据挖掘工具评价标准,主要选择聚类分析和关联规则两个数据挖掘模型重点研究。论文对不同高速公路事故多发点鉴定方法对比分析,最终确定选择数据挖掘中的聚类算法来对路段进行事故多发点辨别。以高速公路路段K885+200到K928+200作为案例基础,采用累计频率曲线法对此路段进行事故多发点初步确定,再采用DBSCAN空间聚类分析法通过MATLAB实现结果修正,进一步确定准确事故多发点。论文在所研究高速公路路段的事故多发点输出结果基础上,运用关联规则挖掘方法进行事故影响要素确定,选择适用于单条高速公路的Apriori算法,通过MATLAB得到满足于最小支持度和最小置信度的高速公路交通安全影响要素。本研究的主要成果可为高速公路交通安全的设计和管理提供一定的技术支持和理论参考,对低等级公路的安全管理同样具有借鉴作用。
[Abstract]:In 13th Five-Year, China's planning program has put forward new requirements for the transportation industry: by 2020, a safe, convenient, efficient and green modern comprehensive transportation system has been built. At present, the number of total mileage in the highway is the first in the world, the number of motor vehicles also increases greatly, and the traffic safety is becoming more and more prominent. Road traffic safety has become more and more obvious. On the basis of the statistical analysis of the occurrence of a highway accident in Shandong, the paper summarizes the characteristics of the highway traffic accident, mainly from the time characteristics, spatial characteristics, climate distribution characteristics and morphological characteristics of the traffic safety accidents. On this basis, the influence factors of highway traffic safety are analyzed. The multi source heterogeneity of traffic safety data of regional highway network is analyzed, and the data fusion method is used to preprocess the traffic safety data. On this basis, the process of data mining to obtain information is studied, and a unified evaluation standard for data mining tools is established. Two data mining models of cluster analysis and association rules are selected to focus on the research. This paper compares and analyzes the multiple point identification methods of different Expressway accidents, and finally determines the clustering algorithm in the selection of data mining to identify the accident multiple points. Taking the highway section K885+ 200 to K928+200 as the case basis, the cumulative frequency is adopted. The rate curve method is preliminarily determined for the accident multiple points in this section, and then the DBSCAN spatial clustering analysis method is used to correct the results through MATLAB to further determine the multiple point of accurate accident. The Apriori algorithm suitable for single highway is selected and the influence factors of expressway traffic safety are satisfied with minimum support and minimum confidence through MATLAB. The main results of this study can provide technical support and theoretical reference for the design and management of highway traffic safety, and the safety management of low grade highway. It can also be used for reference.

【学位授予单位】：山东科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.31

【参考文献】