基于时间序列算法的城市轨道交通客流量预测

Passenger Flow Prediction of Urban Rail Transit Based on Time Series Algorithm

摘要: 城市轨道交通客流量预测是城市轨道交通网络研究的重要内容，对现有已开通线路的历史客流量进行建模预测，能为运管部门提供准确的客流变化信息，优化列车运营计划，也能为站点在突发情况下的异常客流提供应对策略。在对地铁车站客流数据处理的基础上，建立了基于自回归移动平均模型(ARIMA)和Prophet的城市轨道交通客流预测模型，并比较了两种模型的预测精度。结果表明：Prophet模型在预测精度上相对高于ARIMA模型，但两种单项模型的预测情况仍不够准确。因此，在后续的客流预测研究中，应结合多种模型进行客流预测，以及根据不同时间粒度进行客流预测。

Abstract: Urban rail transit passenger flow forecasting is an important content of urban rail transit network research. Modeling and forecasting the historical passenger flow of existing opened lines can not only provide accurate passenger flow change information for the transportation management department, optimize the train operation plan, but also provide countermeasures for the abnormal passenger flow of the station in case of emergencies. Based on the data processing of subway station passenger flow, an urban rail transit passenger flow prediction model based on autoregressive moving average model (ARIMA) and prophet is established, and the prediction accuracy of the two models is compared. The results show that the prediction accuracy of Prophet model is relatively higher than ARIMA model, but the prediction of the two single models is still not accurate enough. Therefore, in the follow-up study of passenger flow forecasting, passenger flow forecasting should be combined with various models and according to different time granularity.