王昕,王玥,袁柯楠

• 1:北京信息科技大学理学院

摘要(Abstract)：

运用数据挖掘技术进行城市交通网络的多模式流量预测及影响因素分析，对于提高交通系统的效率和安全性，支持城市的可持续发展具有重要意义。提出一种基于轻量级梯度提升机（light gradient boosting machine, LightGBM）的多模式交通流量预测算法。根据历史流量数据及多种天气因素，使用灰色关联分析（grey relation analysis, GRA）和Shapley加性解释（Shapley additive explanation, SHAP）对不同交通模式下的天气特征进行筛选，完成城市交通网络中铁路、公交车等6种模式交通流量的鲁棒性预测。仿真试验结果显示，除民航外，GRA-LightGBM组合模型的预测精度在其余5种交通模式的流量预测中均优于极端梯度提升（extreme gradient boosting,XGBoost）模型、支持向量回归（support vector regression, SVR）模型和差分自回归移动平均（autoregressive integrated moving average, ARIMA）模型，表明GRA-LightGBM组合模型兼具时序感知和天气特征融合能力。

关键词(KeyWords)： 多模式;数据平滑;灰色关联分析;轻量级梯度提升机;交通流量预测

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目（71501016）

作者(Author): 王昕,王玥,袁柯楠

DOI: 10.16508/j.cnki.11-5866/n.2025.03.006

参考文献(References)：

• [1]王丽娜.基于历史客流数据的城市轨道交通网络客流量预测及调度研究[D].北京：北京交通大学，2011.WANG L N. Traffic forecasting and scheduling of the urban passenger rail network based on historical data[D]. Beijing:Beijing Jiaotong University,2011.(in Chinese)
• [2]董升伟.基于改进BP神经网络的轨道交通短时客流预测方法研究[D].北京：北京交通大学，2013.DONG S W. The research of short-time passenger flow forecasting based on improved BP neural network in urban rail transit[D]. Beijing:Beijing Jiaotong University,2013.(in Chinese)
• [3]袁坚，王鹏，王钺，等.基于时空特征的城市轨道交通客流量预测方法[J].北京交通大学学报，2017,41(6):42-48.YUAN J,WANG P,WANG Y,et al. A passenger volume prediction method based on temporal and spatial characteristics for urban rail transit[J]. Journal of Beijing Jiaotong University,2017,41(6):42-48.(in Chinese)
• [4]黄海超，陈景雅，王爽，等.多因素轨道交通客流量预测模型研究[J].华东交通大学学报，2021,38(3):61-66.HUANG H C,CHEN J Y,WANG S,et al. Multi-factor rail transit passenger flow prediction model[J]. Journal of East China Jiaotong University,2021,38(3):61-66.(in Chinese)
• [5]左婷.多模式交通条件下城市道路动态网络均衡研究[D].北京：北京交通大学，2013.ZUO T. Study on dynamic network equilibrium for urban road in multimodal traffic[D]. Beijing:Beijing Jiaotong University,2013.(in Chinese)
• [6]李烨.城市多方式交通网络客流分配问题研究[D].南京：东南大学，2018.LI Y. The research of traffic assignment problem on urban multimodal network[D]. Nanjing:Southeast University,2018.(in Chinese)
• [7]葛伟.多模式混合交通网络用户均衡研究[D].南京：南京大学，2021.GE W. Research on multimodal user equilibriumon mixed traffic network[D]. Nanjing:Nanjing University,2021.(in Chinese)
• [8]陈峻.城市多模式交通网络与广义枢纽协同优化研究综述[J].交通科技与经济，2023,25(2):1-10.CHEN J. Collaborative optimization of multi-modal transportation networks and generalized hubs:a review[J].Technology&Economy in Areas of Communications,2023,25(2):1-10.(in Chinese)
• [9]邓润飞，白桦，刘秀彩.江苏省城际多出行模式客流出行特征分析[J].无线互联科技，2023,20(21):129-132.DENG R F,BAI H,LIU X C. Analysis of passenger flow characteristics of intercity multiple travel modes in Jiangsu province[J]. Wireless Internet Science and Technology,2023,20(21):129-132.(in Chinese)
• [10]黄盛，李卫江，朱梦茹，等.降雨事件对上海地铁通勤客流时空影响的精细尺度研究[J].地球信息科学学报，2022,24(2):249-262.HUANG S,LI W J,ZHU M R,et al. A fine-scale study on spatio-temporal patterns of metro commuter flows under rainfall events in Shanghai[J]. Journal of Geo-information Science,2022,24(2):249-262.(in Chinese)
• [11]王殿海，黄莉莎，曾佳棋，等.降雨对城市路网交通状态的影响机理[J].交通运输系统工程与信息，2023,23(3):48-55.WANG D H,HUANG L S,ZENG J Q,et al. Influence mechanism of rainfall on traffic state of urban road network[J].Journal of Transportation Systems Engineering and Information Technology,2023,23(3):48-55.(in Chinese)
• [12]高维.降雨扰动下城市多模式交通系统韧性度量与预测研究[D].哈尔滨：哈尔滨工业大学，2024.GAO W. Research on resilience measurement and prediction of urban multi-modal transportation systems under rainfall disturbances[D]. Harbin:Harbin Institute of Technology,2024.(in Chinese)
• [13]寇敏.融合天气特征的高速公路交通流预测研究[D].济南：山东交通学院，2024.KOU M. Research on highway traffic flow prediction based on weather characteristics[D]. Jinan:Shandong Jiaotong University,2024.(in Chinese)
• [14]寇敏，张萌萌，赵军学，等.融合天气特征的WOA-GRU高速公路短时交通流预测[J].公路，2024,69(4):193-199.KOU M,ZHANG M M,ZHAO J X,et al. Short-term highway traffic flow prediction using WOA-GRU model with fusion of weather features[J]. Highway,2024,69(4):193-199.(in Chinese)
• [15]邢雪，尹子赫，万乐.结合多变量气象因素的共享单车需求预测方法[J].智能计算机与应用，2025,15(1):178-186.XING X,YIN Z H,WAN L. Demand forecasting method of shared bikes combined with multivariate meteorological factors[J]. Intelligent Computer and Applications,2025,15(1):178-186.(in Chinese)