基于CART的驾驶员路径选择模型研究
首发时间:2015-12-15
摘要:针对驾驶员路径选择行为估计误差较大的问题,在SP问卷调查的基础上,基于分类树方法设计了一种驾驶员路径选择模型,并对分类树方法中的剪枝效益函数进行了改进。实证分析结果表明,分类树方法可以有效地解决驾驶员路径选择行为估计问题,其估计误差明显低于已有的Probit模型和Logit模型,而且所设计的剪枝效益函数,可以进一步改善驾驶员路径选择行为估计的效果。
关键词: 交通运输规划与管理; 动态交通引导; 驾驶员路径选择; SP调查
For information in English, please click here
Drive Route Choice Model based on CART
Abstract:In this paper a survy about driver route choice behavior is conducted based on Stated Preference method and a driver route choice model is proposed based on the Classification And Regression Tree method aiming at the high estimation error of driver route choice behavior. Also a new pruning benefit function is built for the CART model. The empirical results show that the Classification And Regression Tree method can be used to estimate the behavior of driver route choice and its estimate error is obviously lower than Probit model and Logit model. Also, the proposed pruning benefit function can further enhance the estimate effect of driver route choice behavior.
Keywords: transportation planning and management dynamic traffic guidance driver route choice;stated preference survy
论文图表:
引用
No.4667804111963214****
同行评议
共计0人参与
勘误表
基于CART的驾驶员路径选择模型研究
评论
全部评论0/1000