摘要
为解决传统的道路黑点鉴别方法中存在的小样本问题及事故样本的信息损失问题,深入分析了事故数据的小样本特性及事故样本反映路段行车风险的多维性;并引入信息分配技术对交通事故数据进行处理,通过计算和比较事故样本分配给各路段单元的事故信息量(而不是事故频次)进行道路黑点鉴别。结果表明:信息分配技术能充分挖掘小样本事故数据提供的道路黑点鉴别信息,事故信息量比事故频次更能反映各路段单元行车风险的大小。
In order to solve small sample problem and reduce information loss of accident samples in traditional road black spot identification methods, small sample characteristics of accident data and multidimensionality of road risk indicated by accident sample were deeply analyzed, then, information assignment technology was introduced to process traffic accident data. Road black spots were identified by comparison of accident information volume instead of accident frequency in each road section element. Results show that information assignment technology can make full use of road black spot identification information supported by accident data of small sample, and accident information volume has better reflection of ride risk in each road section element than that of accident frequency.
出处
《中国公路学报》
EI
CAS
CSCD
北大核心
2007年第4期122-126,共5页
China Journal of Highway and Transport
基金
国家西部交通建设科技项目(200431822333)
上海市科学技术委员会科研计划项目(042112015)
关键词
交通工程
道路安全
信息分配技术
黑点鉴别
小样本
事故信息量
事故频次
traffic engineering road safety
information assignment technology
black spot identification small sample
accident information volume
accident frequency