基于波动聚集性的城际高铁客流量预测  

Predicting intercity high-speed railway passenger flow based on volatility clustering

作　　者：耿立艳 鲁荣利 李新杰 GENG Liyan;LU Rongli;LI Xinjie(School of Economics and Management,Shijiazhuang Tiedao University,Shijiazhuang 050043,China;Hebei Qugang Expressway Development Co.,Ltd,Baoding 071000,China)

机构地区：[1]石家庄铁道大学经济管理学院,河北石家庄050043 [2]河北曲港高速公路开发有限公司,河北保定071000

出　　处：《铁道科学与工程学报》2019年第8期1890-1896,共7页Journal of Railway Science and Engineering

基　　金：国家自然科学基金青年基金资助项目(61503261);河北省交通运输厅科技计划资助项目(QG2018-4);河北省高等学校青年拔尖人才计划资助项目(BJ2014097)

摘　　要：由于受到众多因素的影响,城际高铁客流量序列呈现出波动聚集性特征,常用的预测方法很难准确揭示这种波动聚集性特征,一定程度上限制了城际高铁客流量预测精度的提高。为解决该问题,将自回归差分移动平均(ARIMA)模型与广义自回归条件异方差(GARCH)模型相结合,提出城际高铁客流量的ARIMA-GARCH预测模型。先构建城际高铁客流量序列的ARIMA模型,再利用GARCH模型刻画ARIMA模型残差的波动聚集性。利用某车站的城际高铁客流量数据检验ARIMA-GARCH模型的有效性。研究结果表明:ARIMA-GARCH模型刻画出了城际高铁客流量的波动聚集性特征,其短期、中期、长期预测精度均高于ARIMA模型。随着预测步数的增加,ARIMA-GARCH模型的预测精度逐渐下降。Intercity high-speed railway passenger flow sequence is characterized by volatility clustering due to the influence of many factors.It is difficult for commonly used prediction methods to reveal the volatility clustering accurately,which limit the improvement of the precision to predict intercity high-speed railway passenger flow.To solve this problem,autoregressive integrated moving average (ARIMA) model and generalized autoregressive conditional heteroscedasticity (GARCH) model (ARIMA-GARCH) were proposed in this paper to predict intercity high-speed railway passenger flow.First,ARIMA model was constructed by using intercity high-speed railway passenger flow sequence.Then,GARCH model was used to describe the volatility clustering in the residual from the ARIMA model.The effectiveness of the proposed model was tested based on the intercity high-speed railway passenger flow data from a station.Results show that the ARIMA-GARCH model captures the volatility clustering in the intercity high-speed railway passenger flow.The precision from the ARIMA-GARCH model is higher than the ARIMA model in short-,medium- and long-term prediction.With the increase of prediction steps,the prediction accuracy of the ARIMA-GARCH model decreases gradually.

关 键 词：铁路运输 城际高铁 客流量 预测 ARIMA-GARCH模型 

分 类 号：U412.366[交通运输工程—道路与铁道工程]