基于ARIMA算法的地铁乘客流量预测  

Prediction of Subway Passenger Flow Based on ARIMA Algorithm

作　　者：海玲[1] 刘文[1] 刘岩[1] 谷峥[1] 刘智勇[1] HAI Ling;LIU Wen;LIU Yan;GU Zheng;LIU Zhiyong(Artificial Intelligence and Intelligent Mine Engineering Technology Center,Xinjiang Institute of Engineering,Urumqi 830023)

机构地区：[1]新疆工程学院人工智能与智慧矿山工程技术中心,乌鲁木齐830023

出　　处：《计算机与数字工程》2025年第3期666-670,共5页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:61962058);新疆维吾尔自治区自然科学基金面上项目(编号:2019D01A30);新疆维吾尔自治区自然科学基金项目(编号:2020D01B20)资助。

摘　　要：随着城市人口的日渐增加,带来的突出问题就是地铁线路运输的客流量激增,导致地铁的承载压力变大,给地铁管理部门的运营调度工作带来极大的挑战,针对上述问题,急需一种地铁乘客流量预测方法来解决地铁运管部门运营调度的难题。基于此,论文在时间序列法预测地铁乘客流量的基础上,引用了ARIMA模型,基于数据分析筛选,通过对地铁客流历史数据的特征变化分析,进行了20条站点线路的数据稳定性优化及白噪声检验,使用自相关和偏相关图来对模型参数估值,最后测试ARIMA模型的拟合度,对20个站点线路进行预测,分析地铁的客流数据变化情况,从而得到一个模拟计算后的客观预测数据,为地铁运营调度部门提供科学决策。With the increase in urban population,the outstanding problem is the traffic line transportation soared,leading to the subway of bearing pressure,metro operation dispatching work to bring huge challenges,according to the above problems,it is badly in need of a subway passenger flow prediction method to solve the problem of underground pipe department operation scheduling.Based on this,this article in time series method to predict the metro passenger flow,on the basis of quoting the ARIMA model,based on the data analysis,screening,through the analysis of the characteristic changes of subway traffic history data,the data stability of the 20 sites routes optimization and white noise test,the autocorrelation and partial correlation chart is used to valuation of model parameter,Finally,the fitting degree of ARIMA model is tested to predict the lines of 20 stations and analyze the changes of subway passenger flow data,so as to obtain an objective forecast data after simulation calculation,which provides scientific decision-making for the subway operation and dispatching department.

关 键 词：数据分析 乘客流量 ARIMA模型 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]