Research On High-speed Railway Passenger Flow Forecast Based On Influencing Factors Analysis

For the railway industry,passenger flow forecasting is the prerequisite for route planning and design,the basis for transportation resource allocation,and the basis for daily marketing decisions.Accurate passenger flow prediction helps them to allocate resources reasonably and improve operating revenue.China’s high-speed railway has been in a rapid development stage in recent years,and a large number of high-speed rail lines are under construction.For these newly built high-speed rail lines,because of the lack of historical data,it is difficult to grasp the law of passenger flow,and the difficulty of predicting passenger flow is greatly increased.At the same time,for some high-speed rail lines that have been opened,due to changes in external conditions related to market competition,the historical data may also affect the reference value of the passenger flow forecast.Therefore,how to predict the passenger flow of HSR more accurately based on various factors that determine passenger flow under the premise of lack of historical data is one of the problems faced by the development and operation of HSR,and is also the problem that this paper tries to solve.This paper first analyzes the fluctuation law of HSR passenger flow and its related factors,clarifies the possible influencing factors of HSR passenger flow,discusses the characteristic representation methods of various influencing factors and constructs the feature variable candidate set.Secondly,it reviews the commonly used high-speed rail passenger flow prediction methods and focuses on the e-SVR algorithm-based high-speed rail passenger flow prediction model used in this paper.On this basis,using the idea of control variables,by arranging and combining the characteristic variables,32 models were constructed,and by substituting the training set data into the model,the training results were analyzed to obtain the optimized high-speed rail passenger flow prediction model.Finally,the optimized prediction model obtained in the previous step is used to predict the passenger flow in the test set,and compared with the model before optimization and the main prediction methods currently in practical use to verify the effectiveness of the model.34 figures,18 tables,47 references.