准确的客运车辆到站预测是城市智慧交通的基础服务,有助于减少信息盲区,优化车辆运营调度.提出了一种基于SVM的到站预测模型,考虑道路因素、大型节假日、天气、路况、运行距离、运行时间、排班信息七个因素的影响,改进道路路段为道路类型因素,使模型更适合于客运车辆.在此基础上,用遗传算法做参数寻优提高模型训练效率.以深圳-广州的客运班车GPS数据完成实验,对比证明SVM+GA模型应用于客运车辆行程时间预测具有更好的适应客性,准确高效.%Accurate coach arrival time prediction is one of the infrastructure services in intelligent urban transportation,which helps reducing information blind-spots and optimizating coach bus schedule. An arrival time prediction model for coach bus is pro?posed. It has 7 features including road factors,holidays,weather,road conditions,distance,time,scheduling information. And it change straditional feature road segments to road type factor,in order to make the model more suitable for coach. Besides,this pa?per uses genetic algorithm to find model 's optimal parameters. The experimental results of coach bus from Shenzhen to Guangzhou show that the proposed model is more suitable to predict the coach arrival time with higher prediction accuracy.
展开▼
