Passenger Flow Forecast Of Urban Rail Transit Stations Based On Multi-mode Influence Area And Attraction Intensity

In the context of the rapid development of urban spatial planning and big data technology,refined city management has new requirements for metro passenger flow forecast.How to establish a fast response,convenient and accurate direct passenger flow prediction model has become an urgent problem to be solved.From the perspective of space,this paper uses big data analysis methods and spatial modeling techniques to study the direct prediction method and model of metro stations passenger flow based on multi-mode influence area and attraction intensity.The main contents are as follows:(1)The characteristics of rail transit passenger flow and travel requirements are analyzed,and the influencing factors and functions of rail transit passenger flow are analyzed according to the process of rail travel.The key issues of rail transit passenger flow prediction from a spatial perspective are extracted,including the acquisition and analysis of spatial data,Multi-mode influence area and attraction intensity research,passenger flow prediction modeling from the perspective of space,and formed corresponding research ideas and framework.(2)Acquire data according to the sources and characteristics of rail transit-related spatial data,characterize and quantify the spatial elements affecting station passenger flow,and achieve multi-source data fusion and integration based on Jaro-Winkler algorithm and spatial superposition method to improve the availability and standardization of data.Multi-scale analysis of the relevant spatial data obtained by the process will lay a data foundation for model construction.(3)Based on the distribution characteristics of spatial elements and the characteristics of the available data,the direct influence area is studied using the threshold-based Amap lattice distribution method,and the gray distance decay model based on spatial data sampling is used to study the indirect influence area,and the corresponding algorithm is given.On this basis,the multi-mode attraction intensity model was constructed and calculated,which provides conditions for the establishment of passenger flow direct prediction model.(4)Considering the heterogeneity and dependence of passenger flow influencing factors at different spatial levels,a direct predictive modeling method for station passenger flow with comprehensive consideration of global characteristics and local characteristics is proposed.Based on the spatial integration data,the original indicators of passenger flow influencing factors were constructed,and the original indicators were reconstructed in combination with the multi-mode attraction intensity model of the station.The spatial characteristics and data characteristics of the reconstruction index are analyzed.Based on this,the selection of characteristic variables is carried out.The MGWR model of passenger flow prediction from a double-layer space perspective is constructed,and the prediction effect is compared with the GR model and the GWR model.(5)Taking Beijing metro as a case to verify and analyze the research methods in this paper.