基于ARIMA-Kalman滤波混合算法的铁路进站客流预测方法

Prediction Method of Railway Station Passenger Flow Based on ARIMA-Kalman Filter Hybrid Algorithm

轨道交通车站客流预测,是优化车站客运组织、提高运营安全和运输效率的有效途径。针对传统ARIMA模型对客流量预测性能较差的问题,提出一种基于ARIMA-Kalman滤波混合预测方法。具体通过建立ARIMA模型训练实验样本,结合Kalman滤波器,建立预测递推方程,最终利用Kalman滤波预测方法对客流量进行预测。基于江门东站进站客流数据的仿真实验结果表明,相较于单一ARIMA模型,所提出的ARIMA-Kalman滤波混合算法预测结果的均方根误差降低了257.106,平均绝对误差降低了145.675,平均绝对百分比误差下降了5.655%,证明了所提出的混合算法预测精度更高。

Passenger flow forecasting at railway stations is an effective way to optimize station passenger organization and improve operational safety and transport efficiency.Aimed at the shortage of poor performance of the traditional ARIMA model for passenger flow prediction,a hybrid prediction method based on the ARIMA-Kalman filter is proposed.Specifically,the ARIMA-Kalman filter hybrid forecasting method is used to forecast passenger flow by building ARIMA model training experimental samples,combining the Kalman filter,and establishing forecast recursive equations.The results of the simulation experiments based on the inbound passenger flow data of Jiangmen East Station show that,compared with the single ARIMA model,the root mean square error of the proposed hybrid ARIMA-Kalman filter prediction algorithm is reduced by 257.106,the average absolute error is reduced by 145.675 and the average absolute percentage error is reduced by 5.655%,which proves that the prediction accuracy of the proposed hybrid algorithm is higher.