摘要
道路交通安全水平的重要标志就是道路交通事故发生量.为解决当前交通事故量预测精度不高、时间拐点数据预测效果差的问题,以及在交通管理系统中提供更加准确的预测数据帮助交通管理部门做出科学的决策,本文针对我国年周期交通事故建立了基于GBRT(Gradient Boosted Regression Tree)的交通事故模型.通过训练交通事故相关数据对未来交通事故死亡人数进行预测,并与多种回归模型、神经网络模型进行对比实验,结果显示GBRT模型具有拟合效果佳、训练时间短、高鲁棒性的优势,能够更准确、高效的对交通事故安全水平进行预测.
The important manifestation of road traffic safety level is the amount of road traffic accidents.In view of the problem that the current traffic accident volume prediction accuracy is not high,the time inflection point data prediction effect is poor,and provide more accurate prediction data to the traffic management department,help make scientific decisions in the traffic management system,this paper established a traffic accident model based on GBRT(Gradient Boosted Regression Tree)for China’s annual traffic accidents,and predicts the number of traffic accident deaths in China by training traffic accident related data.Compared with various regression models and neural network models,the results show that the GBRT model has the advantages of best fitting effect,short training time and high robustness,which can predict traffic accident safety level more accurately and efficiently.
作者
杨文忠
张志豪
柴亚闯
温杰彬
杨蒙蒙
富雅玲
YANG Wenzhong;ZHANG Zhihao;CHAI Yachuang;WENG Jiebin;YANG Mengmeng;FU Yaling(School of Information Science and Engineering,Xinjiang University,Urumqi Xinjiang 830046,China)
出处
《新疆大学学报（自然科学版）》
CAS
2020年第1期36-43,共8页
Journal of Xinjiang University(Natural Science Edition)
基金
新疆维吾尔自治区自然科学基金项目“高速公路VANET预警信息广播传输机制研究”(2017D01C042).
关键词
交通事故
集成学习
GBRT
预测
回归
traffic accident
ensemblelearning
GBRT
prediction
regression
