基于SVM的铁路货运站装车数预测  

Prediction on Car Loading Quantities of Railway Freight Stations Based on SVM

作　　者：余姣姣 Yu Jiaojiao(School of Traffic & Transportation,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学交通运输学院,甘肃兰州730070

出　　处：《物流技术》2019年第3期38-42,共5页Logistics Technology

摘　　要：铁路货运站装车数作为铁路货运量的组成部分,对其的预测具有重要意义。采用相空间重构技术对一维时间序列进行重构,将重构的多维数据作为支持向量机的输入,并利用粒子群算法对支持向量机的参数进行优化,运用得到的最优参数训练模型。用此模型对广铁(集团)公司的岳阳北、湘潭东、下元等货运站的装车数进行预测,预测结果表明:该方法的预测精度明显比灰色预测好;对数据重构的嵌入维数和时间延迟导致预测精度的不同,岳阳北和湘潭东的预测效果较好,而下元的预测误差较大。In this paper,we reconstruct the one-dimensional time series using the phase-space reconstruction technique,use the multi-dimensional data reconstructed as input to the support vector machine,optimize the parameters of the SVM using the particle swarm optimization(PSO),and train it using the optimal parameters obtained.Next,we apply the process to predicting the car loading quantities of a few freight stations of the Guangzhou Railway(Group)Company,including the Yueyang North Station,Xiangtan East Railway Station,and Xiayuan Railway Station,etc.The result of the prediction shows that the accuracy of the process is obviously better than that of the gray prediction model;and the difference in embedding dimension and time delay in data reconstruction would affect the accuracy of the prediction:the prediction concerning Yueyang North Station and Xiangtan East Railway Station is more true to reality than that of Xiayuan Railway Station.

关 键 词：铁路货运站 装车数 相空间重构 粒子群算法 支持向量机 

分 类 号：F532[经济管理—产业经济]