Study On Short-time Passenger Flow Forecasting Based On Rail Transit Network

Urban rail transit has become an important part of modern urban transport system.With the increasing speed of urban rail transit network construction,the impact of rail transit on passenger traffic is increasing,and the real-time characteristics of passenger traffic are becoming more and more obvious.Therefore,it is necessary to analyze the historical passenger flow data and then to establish the practical passenger flow forecasting model to predict the short-time passenger flow.The prediction of passenger flow information plays an important role in the establishment of a reasonable operation scheduling plan for the operation department.The sample data in this paper is obtained through the automatic fare collection system in urban rail transit passenger flow.According to the characteristics of the urban rail transit passenger flow,combined with the current prediction methods in urban rail transit passenger flow,we apply the short-term prediction theory to the Shanghai urban rail transit system,and carry on the related research.The main work is as follows.1.The time,section,station nonequilibrium factor are used to describe the nonequilibrium characteristic.According to the nonequilibrium factor of different stations,the stations are divided into three types: the very unbalanced type,the unbalanced type and the balanced type.2.The time series model,support vector machine model and combined forecasting model are studied in depth.And the moving average method of adaptive window width and the wavelet threshold denoising combination support vector machine are proposed for short-time passenger flow prediction.We use the Shanghai urban rail transit passenger flow data and the standard evaluation method to analyze the prediction results.The results show that the wavelet threshold denoising combination of support vector machine has higher prediction accuracy in predicting the very unbalanced stations and the unbalanced stations;and the moving average method has the advantages of better prediction accuracy and shorter time in predicting the balanced stations.3.By using the Matlab and JAVA development and design of graphical user interface program,we realize the prediction and analysis of passenger flow in the road network,achive the functions of the visualization of passenger flow and complete the display of the research content.