A mathematical model describing the risks of road accidents has been built on the basis of statistical data of drivers’ accident rate. It has been revealed that drivers can be divided by the degree of their accident ...A mathematical model describing the risks of road accidents has been built on the basis of statistical data of drivers’ accident rate. It has been revealed that drivers can be divided by the degree of their accident proneness into four categories with sharply differing probabilities of road accidents. It has been shown that there is a possibility of classification of drivers in accordance with specified categories.展开更多
道路交通事故多发点事故发生频率高且严重性突出,为提高道路通行的安全与效率,需要找到事故多发点所在位置。针对现有密度聚类算法对交通事故多发点识别时需要设置中心点个数以及容易扩大聚类范围等问题,提出一种限制簇扩展的自适应搜...道路交通事故多发点事故发生频率高且严重性突出,为提高道路通行的安全与效率,需要找到事故多发点所在位置。针对现有密度聚类算法对交通事故多发点识别时需要设置中心点个数以及容易扩大聚类范围等问题,提出一种限制簇扩展的自适应搜索密度峰值聚类算法(limit cluster expansion and adaptive search clustering by fast search and find of density peaks,LA-CFDP)。LA-CFDP算法通过增加中心点限制条件自动确定中心点个数,引入参数扩展因子限制簇扩展范围,从而提高算法对事故多发点识别的适应性和准确性。在英国四个城市2019年交通事故数据集上的实例分析表明,LA-CFDP算法对四个城市聚类结果的轮廓系数值达到0.72~0.92,DBI值均降低到0.37以下。聚类结果符合事故多发点的定义及特征,能够为交通事故多发点治理提供可靠依据。展开更多
文摘A mathematical model describing the risks of road accidents has been built on the basis of statistical data of drivers’ accident rate. It has been revealed that drivers can be divided by the degree of their accident proneness into four categories with sharply differing probabilities of road accidents. It has been shown that there is a possibility of classification of drivers in accordance with specified categories.
文摘道路交通事故多发点事故发生频率高且严重性突出,为提高道路通行的安全与效率,需要找到事故多发点所在位置。针对现有密度聚类算法对交通事故多发点识别时需要设置中心点个数以及容易扩大聚类范围等问题,提出一种限制簇扩展的自适应搜索密度峰值聚类算法(limit cluster expansion and adaptive search clustering by fast search and find of density peaks,LA-CFDP)。LA-CFDP算法通过增加中心点限制条件自动确定中心点个数,引入参数扩展因子限制簇扩展范围,从而提高算法对事故多发点识别的适应性和准确性。在英国四个城市2019年交通事故数据集上的实例分析表明,LA-CFDP算法对四个城市聚类结果的轮廓系数值达到0.72~0.92,DBI值均降低到0.37以下。聚类结果符合事故多发点的定义及特征,能够为交通事故多发点治理提供可靠依据。