Study On Prediction Of Urban Subway Passenger Flow Based On OD Matrix

With the rapid development of China’s economy and the continuous expansion of the city scale,the demand of residents for public transport continues to grow,and problems such as traffic congestion and traffic pollution are increasingly prominent However,with the expansion of the scale of subway operation network and the complexity of operation mode.The main work and conclusions are as followsFirst of all,this paper analyzes the current situation of traffic congestion,analyzes the current problems such as large passenger flow and congestion.And compares the research status and methods at home and abroad.According to the data characteristics of short-term passenger flow of metro,the effectiveness of the original data is tested,the passenger flow data is extracted,the traffic card data processing system is established,the traffic card data processing process is designedThen,this paper will use linear regression,support vector regression and long-term memory model to predict the short-term passenger flow of subway station The attributes used in these models include(1)time attributes extracted from the passenger flow sequence itself,representing its autocorrelation characteristics;(2)spatial attributes extracted from OD matrix,representing the impact of other stations on the passenger flow of the target station;(3)environmental attributes obtained from external websites,representing the impact of external environment on the travel behavior.Finally,the OD estimation method and passenger flow prediction model are applied to the subway data of a city,and compared with the classic time series prediction model Arima.The experimental results show that:(1)the long-term and short-term memory models and ARIMA models perform well without considering the spatial attributes;(2)the spatial attributes can effectively improve the prediction accuracy of the model.