基于SSA-LSTM的城市轨道交通短时客流预测

Short-term Passenger Flow Prediction of Urban Rail Transit Based on SSA-LSTM

为了减小噪声在城市轨道交通短时客流预测中的干扰,提出了一种基于奇异谱分析(SSA)和长短时记忆网络(LSTM)模型的短时客流预测方法。该模型利用SSA对原始客流数据进行嵌入、分解、分组、重组,重新划分为趋势、周期、残差三部分,其中残差部分即视为噪声,将去噪后的两个部分作为LSTM模型的输入。并利用上海地铁1号线进出站客流数据对模型的有效性进行验证。结果表明,SSA-LSTM混合模型的预测精度更高。

In order to reduce the interference of noise in short-term passenger flow prediction of urban rail transit,a short-term passenger flow prediction method based on singular spectrum analysis(SSA)and long short-term memory network(LSTM)model is proposed.The model uses SSA to embed,decompose,group and reorganize the data,and divides the original passenger flow into three parts:trend,cycle and residual.The two parts after denoising(residual part)are used as the input of LSTM model.Validation of the model based on the passenger flow data of Shanghai Metro Line 1.The original passenger flow collected as input,by adjusting the relevant parameters to predict.Finally,the prediction results are compared with those only using LSTM model.The results show that the SSA-LSTM hybrid model has higher prediction accuracy and smaller error.