| Nowadays,traffic congestion has become a major problem that disturbs the sustainable development of major cities and seriously affects the traffic quality of urban road network.The traditional road congestion evaluation usually takes the road section as the minimum analysis object,and describes the operation status of the road network through the accumulation or average of road sections.However,due to the neglect of the interaction between road sections,the interaction relationship of road network cannot be analyzed scientifically,and it is difficult to evaluate the overall operation quality of road network and improve the operation efficiency.How to scientifically and objectively represent the temporal and spatial characteristics of road network operation and evaluate the operation quality of urban road network has become an important issue in the field of urban traffic operation.This article from the regional level to explore the status of urban road network based on the multi-source urban traffic data and road network dynamic operation data,build the evaluation index system of city regional road network running quality,to reflect the different types of regional operation of multidimensional characteristics,and put forward the regional road network running quality influence model,realize the relationship between the factors of regional road network running state score and characterization.Paper firstly combed the domestic and foreign about urban traffic running state evaluation,influence factors to explore and study the influence model of the status quo,based on the point of interest,the index of road network structure,the regional transportation and road traffic flow data,from the intensity of traffic,congestion degree and the influence scope of three dimensions to build the urban regional road network running state evaluation index system and index calculation method.Taking59 districts in the central city of Beijing as an example,the agglomeration hierarchy method(HAC)was used to cluster the road network running state,and the characteristics of the change of running state in typical districts were analyzed.Based on the classification of the running state of urban regional road network,further quantitative evaluation of the running state of regional road network is carried out.Based on the information entropy theory,a set of road network operation quality scoring algorithm is proposed to quantitatively describe the congestion level of eachregion.In order to further explore the influence characteristics of the running state of regional road network,the influence factor set was established from three aspects:land use attribute,road network structure and traffic flow.By the method of correlation coefficient of regional road network running state influence factors are analyzed in the correlation and the Spearman correlation coefficient analysis,to test the significant influence factors,the results show that office density,density of health care,road network density,high ratio,average daily traffic road,shopping service density,density of sports leisure,residential density,the density of science culture,a total of nine factors will,in turn,on the network running status significantly influence from big to small.Finally,on the basis of the evaluation model and the significance analysis of influencing factors,the paper constructs the influence model of regional traffic operation quality based on XGBOOST algorithm,and applies the optimal model to verify and analyze the accuracy.The comparison experiments with ordinary linear regression,bayesian regression,elastic network regression and support vector machine regression models show that the models constructed in this paper are the best in cross validation and scores,which proves that this model is more suitable for nonlinear and multi-parameter regression analysis.The research conclusion of this paper is helpful for the government and regulatory authorities to identify the key congestion areas,so that the corresponding improvement countermeasures and measures can be made for the areas with different congestion types,which has important practical significance and theoretical support for the exploration and quantitative evaluation of the influencing factors of regional operation quality. |