The Factor Analysis Of Effect On The Transit Ridership Based On Geographically Weighted Regression

The development of public transportation is one of the important measures to solve urban traffic congestion.A reasonable forecast of urban public transportation can give more reasonable bus planning and promote the development of public transportation.Because of the unreasonable socio-economic conditions in different parts of the city,there are also differences in the amount of public transport produced.In order to understand the impact of different socioeconomic factors on the volume of passenger flow generated by the bus,most researchers use the bus station as a research unit to establish a linear regression model to analyze the influencing factors of the bus passenger flow,and have also obtained good results.However,the bus passenger flow data in these studies was obtained through manual investigation and was laborious.In addition,for a large area of study,ordinary linear regression models do not explain the spatial differences in the influencing factors.Therefore,this paper uses bus GPS and IC card swiping data to carry out fusion mining analysis,and obtains the bus passenger flow in different areas of the city.The geo-weighted regression model is used to reveal the spatial heterogeneity of the influencing factors of the bus passenger flow.This paper first reviews the domestic and foreign research results on the influence factors of public traffic flow and the application research of geographically weighted regression model.Then it elaborates on the spatial heterogeneity theory,the multiple linear regression theory and the geographical weighted regression theory in detail,and lays a theoretical foundation for the following research.Then,the process of inferring public traffic using public transportation big data under informatization conditions is introduced in detail,and the bus passenger flow is integrated in the unit of traffic community.Secondly,the dependent variable and independent variable data required for modeling are processed,and the dependent variable and independent variable data are subjected to Log transformation.The spatial autocorrelation and multicollinearity analysis are performed on the independent variable data.In the end,the common linear regression model and GWR model were established to analyze the influencing factors of public traffic flow.The results of the research show that the bus passenger flow inferred from the bus GPS and IC card swiping data is consistent with the actual situation and can represent the amount of bus passenger flow generated in the entire region.The geographically weighted regression model is better than the ordinary linear regression model,and can explain the impact of various influencing factors on local bus traffic.The innovations of the dissertation include the following two points: Firstly,the analysis of the influencing factors of the bus passenger flow is based on the analysis of the bus passenger flow acquisition.The general way of obtaining the bus passenger flow is the bus passenger flow survey.This article uses the bus big data mining to obtain the passenger flow of the bus station and integrates it into the traffic zone Level.Secondly,this paper uses geographic weighted regression model to analyze the influence factors of public traffic passengers at the traffic community level.The results show that the geographically weighted regression model is superior to the ordinary linear regression model in the fitting degree,which is conducive to improving the bus passenger flow forecast in different areas of the city.