Study On Passenger Flow Prediction Of Intercity Passenger Line Based On Improved BP Neural Network

Intercity passenger line is one of the most important modes of transportation between cities.The accurate prediction of passenger flow is the key to improve the efficiency of operating and organizing about intercity passenger lines,which can provide some references for the operation and organization about intercity passenger lines.However,at present,there are few studies concentration on passenger flow prediction about intercity passenger lines,and due to the limitation of historical data and prediction models,the accuracy of passenger flow prediction is unsatisfactory.Based on the characteristics of passenger flow about intercity passenger lines,established non-holiday and holiday passenger flow prediction model,and validated the prediction model based on panel data.The purpose is to improve the accuracy of passenger flow prediction about intercity passenger lines.On the premise of data preprocessing,this paper analyzed the spatial and temporal characteristics of passenger flow about intercity passenger lines.For the spatial characteristics,the passenger flow distribution among nine cities in Guangdong Province about Guangdong-Hong Kong-Macao Greater Bay Area is analyzed.For the temporal characteristics,the passenger flow characteristics of the same intercity passenger line are analyzed from different time scales such as year,quarter,month,week,day and hour.At the same time,the similarities and differences of passenger flow between non-holidays and holidays are emphatically compared.For the study of non-holiday passenger flow forecasting,aimed at the shortage about random selection of initial weights and thresholds in BP neural network,the genetic algorithm based on fitness selection is used to optimize BP neural network,and an improved genetic algorithm optimized BP neural network(IGA-BPNN)for non-holidays passenger flow forecasting model is proposed.In order to verify the accuracy and applicability of the prediction model,after setting the relevant parameters of the model,the non-holiday historical passenger flow data of Shenzhen-Guangzhou intercity passenger line are forecasted and analyzed,the results show that the mean absolute error of the prediction is 6.43%,the prediction accuracy is ideal.At the same time,the passenger flow of nine intercity passenger lines in Guangdong Province about Guangdong-Hong Kong-Macao Greater Bay Area is compared and forecasted,the prediction error is close to that of Shenzhen-Guangzhou intercity passenger line,which proves the applicability of the model.For the study of holiday passenger flow forecasting,classified the different holidays at first,and then analyzed the influence time of different holidays.On the basis of non-holiday passenger flow forecasting model,holiday background passenger flow is forecasted,and by introducing the fluctuation coefficient of holiday passenger flow,the holiday passenger flow forecast model which combined holiday background passenger flow and holiday passenger flow fluctuation coefficient is proposed.The holiday passenger flow data of Shenzhen-Guangzhou intercity passenger line is forecasted and analyzed,the results show that the mean absolute error of the prediction is 6.20%.At the same time,the model is validated by panel data of holiday passenger flow of nine intercity passenger lines in Guangdong Province about Guangdong-Hong Kong-Macao Greater Bay Area,the prediction error is close to that of Shenzhen-Guangzhou intercity passenger lines,which further proved the prediction accuracy and applicability of the non-holiday passenger flow prediction model.