摘要
借助于数学软件MATLAB,运用模糊数学方法,探讨中国交通事故发生及危害的变化规律.首先,作中国交通事故统计数据的散点图,观察发现交通事故的变化同时具有线性增长总趋势和波动性特征,呈现出复杂的非线性组合特点;再对这些数据进行归一化处理,给出交通事故的模糊隶属函数;最后对交通数据进行分布拟合、组合回归处理,得到交通事故非线性模糊分布函数.结果表明,随着时间的变化,中国交通事故的死亡人数呈线性与正态叠加波动上升趋势,受伤人数呈线性与正弦叠加波动上升趋势,发生起数与损失折款都呈双峰正态分布波动趋势,2002年以来交通状况正在好转.得到的结果能客观地反映交通事故的现状分布和精确地刻划中国交通事故发生及危害变化规律,对于交通事故预测、综合治理具有重要的理论意义和应用价值,可供交通事故研究和有关部门制定交通安全管理决策作参考.
The rules of traffic accidents and their hazards in China are studied by virtue of MATLAB and the method of fuzzy mathematics. First, the general trend of linear increase and fluctuant character of the traffic accidents are found by observing the scatter graph of statistical data of traffic accidents in China, which represent a complex nonlinear combinatorial feature. Second, these data are processed by the method of normalization ; fuzzy membership function of traffic accident is arrived at. Finally, a nonlinear fuzzy distribution function of traffic accidents is obtained by way of distribution fitting and combinational regression treatment of the data. The result indicates that with time changing, the death toll in China^s traffic accidents shows the combined upward trend of a linear and Normal distribution fluctuation, the number of the injured shows the upward trend of linear and Sinusoidal fluctuations, accident occurrence and loss both show two-humped Normal distribution. The traffic situation takes a turn for better since 2002. The result can objectively reflect the status distribution and accurately depict the hazard change rule of traffic accidents, which has theoretical significance and application value for forecast and comprehensive control of traffic accidents, thus it can be used as a reference in the research of traffic accidents and decision-making for safety management of concerned departments.
出处
《数学的实践与认识》
CSCD
北大核心
2009年第21期90-95,共6页
Mathematics in Practice and Theory
基金
福建省自然科学基金(2008J0184)
福建省教育厅科技资助项目(JB04116)
关键词
交通事故
隶属函数
数据拟合
组合回归
模糊双峰正态分布
模糊正弦分布
traffic accident
membership function
data fitting
regression combination
fuzzytwo-humped Normal distribution
fuzzy Sinusoidal distribution