铁路车站客流量波动规律与预测方法探讨
Fluctuation Patterns and Prediction Method of Passenger Flow at Railway Stations
任琦璇1
作者信息
- 1. 中国铁路经济规划研究院有限公司 运输研究所,北京 100038
- 折叠
摘要
铁路客运呈现出显著波动特性和复杂变化趋势,准确把握客流变化规律和趋势,实现客流的精准预测,对铁路网络规划、运营组织优化、服务质量提升具有重要意义.以铁路车站客流量2017-2019年历史数据为基础,对节假日和非节假日铁路车站客运量波动规律进行分析,通过日期和节假日属性特征设计,应用提出融合注意力机制的长短时记忆网络(LSTM-AM)模型对铁路某车站日客流量进行预测.结果表明,LSTM-AM模型预测误差小且能够较好地反映出客流变化规律,可以为铁路相关部门提供数据支持.
Abstract
As railway passenger flow shows significant fluctuations and complex trends,to precisely learn the patterns and trends of passenger flow changes and achieve accurate prediction of passenger flow is of great significance for railway network plan-ning,operation organization optimization,and service quality improvement.Based on the historical data of passenger flow at railway stations from 2017 to 2019,this paper analyzes the fluctuation patterns of both holiday and non-holiday passenger flow at stations.By designing date and holiday attribute features,the long short-term memory-attention mechanism(LSTM-AM)model is applied to predict the daily passenger flow at a railway station.The results show that the deviation of prediction results of LSTM-AM model is minor and it can well reflect the changes in passenger flow,which can provide data support for railway operating departments.
关键词
铁路车站/客流量预测/节假日/波动规律/LSTM-AM模型/实证分析Key words
railway station/passenger flow prediction/holidays/fluctuation pattern/LSTM-AM model/empirical analysis引用本文复制引用
出版年
2024
NETL
NSTL
万方数据