基于GRU-Attention组合模型的疫情影响下高速铁路客流预测  

High-speed Rail Passenger Flow Prediction Under the Influence of COVID-19 Based on GRU-Attention Combined Model

作　　者：郑博言 张小强[1,2,3] ZHENG Boyan;ZHANG Xiaoqiang(School of Transportation and Logistics,Southwest Jiaotong University,Chengdu 611756,China;National Engineering Laboratory of Application Technology of Integrated Transportation Big Data,Southwest Jiaotong University,Chengdu 611756,China;National United Engineering Laboratory of Integrated and Intelligent Transportation,Southwest Jiaotong University,Chengdu 611756,China)

机构地区：[1]西南交通大学交通运输与物流学院,四川成都611756 [2]综合交通大数据应用技术国家工程实验室,四川成都611756 [3]综合交通运输智能化国家地方联合工程实验室,四川成都611756

出　　处：《综合运输》2025年第1期93-98,142,共7页China Transportation Review

基　　金：基于旅客出行偏好的高速铁路列车开行方案优化与客运动态价格策略(KYL202112-0200)。

摘　　要：受新冠疫情影响的客流具有更复杂特征与影响因素,为解决受疫情影响客流预测困难的问题,本文将以新冠疫情影响下邕北线高速铁路客流为例,分析了疫情影响下高速铁路客流的特征和影响因素,结合GRU神经网络和注意力机制建立了GRU-Attention组合模型,实现疫情影响下高铁客流的准确预测。将预测结果与SARIMA、SVR、KNN、GRU、LSTM、PSO-GRU、GRU-LSTM等模型进行对比,根据评价指标,GRU-Attention组合模型的MSE、RMSE、MAE和R2值在所有对比模型中最小,预测性能最好。Passenger flow affected by the COVID-19 pandemic exhibits more complex characteristics and influencing factors.In order to address the challenge of predicting passenger flow impacted by the pandemic,this paper takes the passenger flow of the Yongbei high-speed railway line under the influence of the COVID-19 pandemic as a case study.It analyzes the characteristics and influencing factors of high-speed railway passenger flow under the pandemic's influence,and establishes the GRU-Attention combination model by integrating the GRU neural network and attention mechanism to achieve accurate prediction of high-speed rail passenger flow during the pandemic.The prediction results are compared with models such as SARIMA,SVR,KNN,GRU,LSTM,PSO-GRU,and GRU-LSTM.Based on evaluation metrics,the GRU-Attention combination model has the smallest MSE,RMSE,MAE,and the highest R2 values among all compared models,demonstrating the best prediction performance.

关 键 词：高速铁路 客流预测 新冠疫情 特征分析 深度学习模型 注意力机制 

分 类 号：U293.13[交通运输工程—交通运输规划与管理]