【摘要】 提出了一种基于门控循环单元(GRU)神经网络的城市轨道交通短时客流OD(起讫点)预估模型。以实际数据为例,引入同期天气数据,对工作日的数据进行训练预测,并与长短期记忆(LSTM)神经网络模型进行对比。预测结果表明:相对于LSTM模型,GRU模型不仅模型简单、收敛速度明显较快,而且在预估误差和预测稳定性等方面也略优,更适于短时客流OD的快速预测。更多还原

【Abstract】 A short-term passenger origin-destination(OD) forecasting model for urban rail transit based on gated recurrent unit(GRU) neural network is proposed. Based on the practical data, by importing the weather data during the same period, the training prediction of the working day data is conducted and compared with the long short-term memory(LSTM) neural network model. The results show that the convergence speed of GRU is obviously faster than LSTM, the prediction errors and stability are slightly better than LSTM. Therefore, GRU model is more suitable for short-term passenger flow OD prediction.更多还原