基于EMD-SARIMA模型的铁路商品汽车运量预测  

Prediction of Railway Transportation Volume of Commodity Vehicles Based on EMD-SARIMA

作　　者：张旭宁 ZHANG Xuning(Beijing Branch,China Railway Special Cargo Services Co.,Ltd.,Beijing 100071,China)

机构地区：[1]中铁特货物流股份有限公司北京分公司,北京100071

出　　处：《物流技术》2022年第7期87-91,共5页Logistics Technology

基　　金：中国铁路总公司科技研究开发计划课题(N2018J022)。

摘　　要：针对铁路商品汽车月度运量时间序列呈现非平稳、季节性波动等特征,结合经验模态分解(EMD)方法建立了一种新的SARIMA时间序列预测模型。首先利用EMD将时间序列分解成多个相互独立且相互平稳的分量。然后分别对各个分量建立相对应的SARIMA时间序列模型,去除噪声分量。最后进行数据重构,将重构后的数据再进行SARIMA建模,以实现铁路商品汽车月度运量预测。预测结果显示,建立的EMD-SARIMA模型有着很高的预测精度,能够学习获取时间序列铁路商品汽车月度运量的成长过程及发展趋势,挖掘其周期性变化规律,为解决铁路商品汽车物流高质量发展过程中面临的问题起到重要的参考作用。In view of the non-stationary and seasonal fluctuation characteristics of the time series data of the monthly railway transportation volume of commodity vehicles,a new SARIMA time series prediction model is established in combination with the empirical mode decomposition(EMD).First,the time series data is decomposed into multiple independent and stationary components using EMD.Then,the corresponding SARIMA time series model is established for each component,before removing all the noise components.Finally,through data reconstruction,an SARIMA modeling is carried out on the reconstructed data to realize the monthly forecasting of railway transportation volume of commodity vehicles.The forecasting results show that the EMD-SARIMA model established in this paper has a high prediction accuracy,can learn the growth process and development trend of the time series data of the monthly railway transportation volume of commodity vehicles and dig out the hidden change pattern,providing reference for solving the problems faced in the high-quality development of railway transportation of commodity vehicles.

关 键 词：铁路商品汽车运输 运量预测 经验模态分解 SARIMA模型 

分 类 号：U2-9[交通运输工程—道路与铁道工程] F224[经济管理—国民经济]