基于支持向量回归的行程时间预测算法
al control strategy of transportation planning department. For advanced transportation guidance systems， it is a key issue to predict travel times between pairs of points of interest. There are few travel time prediction methods with high probability of prediction error. In this paper， the speed data returned from probe vehicles and microwave radars is used to predict travel times based on support vector regression（SVR）， and the new algorithm is compared to the historical mean algorithm. The experimental results over elevatedroads in Hangzhou show that the SVR based algorithm significantly outperforms the historical mean algorithm.Key words： support vector machine； travel time； intelligent transportation； historical average0 引言行程时间是交通规划、运营和通行能力评估的重要指标。基于预测的行程时间，出行者可以直观地进行路线选择或者出行时间点的选择，交通规划部门能够做出合理的信号控制策略。因此，准确预测行程时间具有重要的应用价值。支持向量机（SVM）[1]是Vapnik在1995年提出的，已经被广泛地应用到监督分类领域。因为该方法采用了结构
<<上一页 下一页>>
广州市越秀区图书馆版权所有。
联系电话：020-87673002
本站访问人数：