基于SVR模型的中国铁路月度货运量预测研究  

Research on Monthly Freight Volume Forecasting of Chinese Railways Based on SVR Model

作　　者：周小丽 ZHOU Xiao-li(School of Management,Zhengzhou University,Zhengzhou 450001,China)

机构地区：[1]郑州大学管理学院,河南郑州450001

出　　处：《物流工程与管理》2024年第10期52-54,58,共4页Logistics Engineering and Management

摘　　要：铁路运输作为综合交通运输体系的一部分,需要优化升级以满足高质量发展的要求。对铁路短期货运量进行预测分析,有助于扩展货运研究领域,进一步了解铁路短期货运现状,明确发展趋势。文中利用中国铁路月度货运量数据(2017-2024年),通过定量和定性的方法对其影响因素进行综合分析,基于最优参数组合的SVR模型进行预测分析,并以MSE和R^(2)作为评价指标评估预测性能。结果表明,基于sigmoid核的SVR模型对铁路货运量有较好的拟合效果。未来研究可深入分析影响因素指标体系构建、评估现行多样化预测模型以及保证预测精度上简化模型等方面。Railway transportation,as part of the comprehensive transportation system,needs optimization and upgrade to meet high-quality development goals.A short-term freight volume predictive analysis of railway was made,which can help broaden research on freight transport,and further improve understanding of current operations,and identify development trends.This study conducts a comprehensive analysis of the factors affecting railway freight volume through both quantitative and qualitative methods,using monthly freight volume data from China Railways(2017-2024).It employs an SVR model with optimal parameter combinations for predictive analysis and uses MSE and R^(2) as evaluation metrics to assess prediction performance.Results show that the SVR model with a sigmoid kernel fits railway freight volume well.Future research could focus on developing an indicator system for influencing factors,assessing existing forecasting models,and simplifying approaches to improve prediction accuracy,and other aspects.

关 键 词：货运量预测 皮尔逊相关分析 SVR 

分 类 号：F532.5[经济管理—产业经济]