Forecasting Method of Urban Rail Transit Ridership at Station level Based on Population Variable in Circle Group

Abstract:

Considering different contribution rates to station riderships (in this essay, it means exit and enter ridership) of population within different distance to the station, it is necessary to classify population into different circle groups according to their distance to the station as a variable of the forecasting model. Population variable in circle group has been certified by partial correlation analysis and at the same time, other significant factors influencing station ridership have been obtained. Because of the irrationality of linear multi variable regression, the back propagation neural networks forecasting model has been built to reflect the high non linear relation between the independent variable and dependent variable. The case study indicates that the forecasting model based on population variable in circle group and BP neural networks significantly precedes other models and meanwhile, it is real time. Based on the above model, the contribution rate model of population in different circle groups to station ridership has been built, where any background variable of the station has been known. The result of the contribution model also indicates the forecasting model based on population variable in circle groups and BP neural network can better reflect the relationship between station ridership and all the factors influencing the riderships.