摘要
为揭示高速公路交通事故发生内在机理、提升高速公路行车安全;首先,基于UW-DriveNet交通大数据平台,提取了2016年美国华盛顿州3万余条交通事故数据,从人、车、路、环境、事故、时间6个维度对数据集进行了样本结构设计;进而,提出了一种考虑定向约束和指标赋权的多维度交互改进Apriori关联规则挖掘算法,以基于区间层次分析法和灰色关联度的主客观联合赋权模型对数据字段进行权重优化;最后,应用该改进算法,对选定的高速公路路段进行了全映射事故致因角度和事故维度自相关角度的多维度交互的关联规则挖掘计算.结果显示,改进的WODMI-Apriori算法能更好地揭示高速公路的事故致因、更精确地识别事故风险因子,其算法精确度较传统Apriori算法提升了82.7%.结果表明,本文提出的WODMI-Apriori算法可作为高速公路交通事故风险识别工作中的一种行之有效的方法,并可为高速公路行车安全水平的提升提供理论指导.
In order to reveal the inner mechanism of freeway crash and improve its driving safety;First,based on the UW-DRIVENet traffic big data platform,more than 30000 traffic crash data in Washington state in 2016 was extracted,the data set was designed with sample structure from six dimensions:people,cars,roads,environment,crash and time.Furthermore,an improved multi-dimensional interactive Apriori algorithm considering directional constraint and index weighting(WODMI-Apriori)was proposed.Weight optimization of data fields was carried out by using the subject-objective joint weight model based on interval analytic hierarchy process and gray relational degree.Finally,the improved algorithm was applied to mining and computing the multi-dimensional interaction of the crash cause and the crash dimension autocorrelation for the selected freeway section.The results show that the improved WODMI-Apriori algorithm can better reveal the crash causes of freeway and identify crash precursors more accurately,and the accuracy of this algorithm is 82.7%higher than that of the conventional algorithm.The results indicate the WODMI-Apriori algorithm proposed in this research can be used as an effective approach for freeway crash risk identification and can provide theoretical guidance for the improvement of freeway safety.
作者
杨洋
袁振洲
王印海
王文成
孙东冶
YANG Yang;YUAN Zhenzhou;WANG Yinhai;WANG Wencheng;SUN Dongye(School of Transportation Science and Engineering,Beihang University,Beijing 100191,China;Department of Civil and Environmental Engineering,University of Washington,Seattle,98195,USA;School of Traffic and Transportation,Beijing Jiaotong University,Beijing 100044,China;Beijing Municipal Institute of City Planning&Design,Beijing 100045,China;National Engineering Laboratory for Transportation Safety and Emergency informatics,China Transport Telecommunications&Information Center,Beijing 100011,China)
出处
《交通工程》
2021年第6期1-10,16,共11页
Journal of Transportation Engineering
基金
中国博士后科学基金资助项目(2021M700333).
