基于季节性差分整合移动平均自回归模型的城市公交短期客流预测

Short Term Passenger Flow Forecast of Urban Public Transport Based on Seasonal Autoregressive Integrated Moving Average Model

为了解决公交车辆过载及空车浪费资源并存问题,提高城市公交服务质量水平,基于公交客流季节性波动及周期性变化特征,构建季节性差分整合移动平均自回归模型,并对城市公交短期客流进行预测;以山东省青岛市K1路公交线路刷卡数据为模型样本,对非平稳的客流时间序列进行1阶7步差分处理,对差分后的数据进行平稳性检验;通过相对信息量计算,确定预测模型中未知参数,对差分处理后的时间序列进行标准化残差检验,检验结果为白噪声序列,得到周期为7的季节性差分整合移动平均自回归预测模型;利用预测模型对2019年7—12月公交客流量进行预测与误差分析。结果表明,模型预测的平均相对误差为4.02%,最大相对误差为8.36%,模型预测精度较高,适用于青岛市公交短期客流量预测。

To solve the problems of public transport vehicles overload and wasting resources of empty vehicles,and improve quality of urban public transport service,a seasonal autoregressive integrated moving average model was constructed based on seasonal fluctuation and periodic change characteristics of public transport passenger flow,and short-term passenger flow of urban public transport was predicted.Taking K1 line bus swipe card data of Qingdao city in Shandong province as model samples,first-order 7-step differential processing was carried out for non-stationary passenger flow time series,and stationarity of difference data was tested.Unknown parameters in the prediction model were determined by calculation of relative information.Standardized residual test was carried out on time series after differential processing.The test results were white noise series,and the seasonal autoregressive integrated moving average model with period 7 was obtained.The forecast model was used to forecast and error analyze passenger flow of public transport from July to December in 2019.The results show that the average relative error of the model prediction is 4.02%,and the maximum relative error is 8.36%.The model has high prediction accuracy,which is suitable for short-term passenger flow prediction of Qingdao public transport.