基于粒子群优化最小二乘支持向量机的交通事故预测方法  

Traffic Crash Prediction Method Using Least Squares Support Vector Machine with Particle Swarm Optimization

作　　者：韦凌翔 赵洪旭[2] 赵鹏飞 钟栋青 陈天昊[2] WEI Lingxiang;ZHAO Hongxu;ZHAO Pengfei;ZHONG Dongqin;CHEN Tianhao(College of Defense Engineering,Army Engineering University of PLA,Nanjing 210007,China;School of material science and Engineering,Yancheng Institute of Technology,Yancheng Jiangsu 224051,China;School of Civil and Transportation Engineering,Beijing University of Civil Engineering and Architecture,Beijing 102616,China)

机构地区：[1]陆军工程大学国防工程学院,南京210007 [2]盐城工学院材料科学与工程学院,江苏盐城224051 [3]北京建筑大学土木与交通工程学院,北京102616

出　　处：《交通工程》2023年第4期94-99,共6页Journal of Transportation Engineering

基　　金：北京市博士后工作经费资助项目(No.2021-zz-111);北京建筑大学青年教师科研能力提升计划资助(No.X21066);江苏省大学生创新训练计划项目;北京建筑大学培育项目专项资金资助(X23044).

摘　　要：为解决交通事故预测中非线性样本影响预测精度的问题,本文构建了基于粒子群算法(PSO)优化的最小二乘支持向量机(LSSVM)的交通事故预测方法.在构建交通事故数LSSVM预测模型的基础上,采用PSO算法优化LSSVM的惩罚系数和核函数宽度;设计了基于粒子群优化最小二乘支持向量机的交通事故预测模型;最后以我国连续48个月的道路交通事故数据建立模型,验证了该预测方法的有效性.实验结果表明:PSO优化LSSVM的交通事故模型比使用经验参数的LSSVM预测模型的预测效果更好.是准确预测交通事故的方法.In order to solve the problem that nonlinear samples affect the prediction accuracy in traffic crash prediction,this paper constructs a traffic crash prediction method based on least squares support vector machine(LSSVM)optimized by particle swarm optimization(PSO).Based on the construction of LSSVM prediction model for traffic crashes,the PSO algorithm is used to optimize the penalty coefficient and kernel function width of LSSVM.A traffic crash prediction model based on particle swarm optimization least squares support vector machine is designed.Finally,a model is established based on road traffic crash data for 48 consecutive months in China,which verifies the effectiveness of the prediction method.Experimental results show that the traffic crash model of PSO optimized LSSVM has a better prediction effect than that of LSSVM prediction model using empirical parameters.It is a method of accurately predicting traffic crashes.

关 键 词：交通安全 交通事故 最小二乘支持向量机(LSSVM) 粒子群优化算法(PSO) 预测模型 

分 类 号：X951[环境科学与工程—安全科学] U491.31[交通运输工程—交通运输规划与管理]