Short-term Passenger Flow Prediction of railway epidemic Based on SARIMA - LSTM combined model (preprint)

ABSTRACT

The huge disturbance caused by the sudden outbreak of COVID-19 to the short-term passenger flow of railway. The daily passenger flow curves of periodic and seasonal non-stationary time series of Spring Festival travel under the COVID-19 pandemic were analyzed, and the combined model based on SARIMA-LTSM was constructed. SARIMA model was used to predict the linear part and LSTM rolling optimization model was used to predict the nonlinear part. Finally, the two prediction results were put into the weighted sum of attention mechanism module, and the GRU gated loop unit was introduced to assist verification. Through case study and analysis, the results show that the prediction results of SARIMA-LTSM combined model have good control and high accuracy, which can provide theoretical basis for the close representation and prediction of short-term passenger flow data set of epidemic emergencies.