Research and Implementation of Temporal Passenger Flow Model

• 1、State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications

Abstract：With the rapid development of urban traffic, the number of passengers has increased dramatically. The passenger travel information offers great value. The method to analyze the passenger flow effectively is particularly important. This paper uses temporal passenger flow model to analyze and predict short-term passenger flow of multiple stations. As the number of stations is huge and is increasing, it takes a lot of manpower to model each station separately. To solve this problem, this paper analyzes the time property of passenger flow by autocorrelation function and other statistical methods, proving that there is a correlation between the current passenger flow and the historical passenger flow. Based on this, we propose a passenger flow cluster algorithm based on temporal factor which measures the similarity of passenger flow between different stations from its trend and numerical point, and clusters the stations. The model can effectively use the time property contained in the passenger flow, aggregate the stations based on similarity of passenger flow and predict passenger flow effectively. Experiment shows that this algorithm has better predictin ability in multiple stations passenger flow prediction problem.

Keywords： computer software and theory passenger flow prediction multiple stations spaticial factor cluster