Modeling And Analysis Of Residents' Travel Behavior Based On Multiple Differences

Today,the disaggregate model is widely used to capture the choice of traffic modes of urban residents,providing a solid theoretical basis and practical basis for the scientific formulation of traffic management policies.However,the terms of model selection,handling travel behavior data,etc.it will have some impact on the construction of disaggregate model of how to analyze and address these impacts in order to establish more accurate disaggregate model to describe the travel behavior of residents to be solved important question.This paper systematically analyzes the differences of travel areas,the differences of different models and the different ways of dealing with missing data,and explores the impact of different kinds of differences on model establishment and policy analysis.This paper firstly uses the 2011 Dalian residents travel survey data to construct a binary Logit model based on different regions(regions,core regions,peripheral regions,peripheral regions-core regions)for comparative analysis.Secondly,based on the core area data of Dalian,the MNL and NL models are established respectively to explore the influence of model differences on commuter travel behavior simulation.Then based on the 2018 Dalian City Ridesourcing car travel behavior survey data,the complete data generated by the non-answer problem is deleted,and the household income of the easily-missing variable is randomly deleted by 10%,30%and 50%become missing data.Comparative analysis with NL,MNL model using a single and multiple data imputation EM,MI imputation missing data and imputation data formed under conditions different miss rate.This paper uses R language programming to construct the disaggregate model.The main effects of the three differences on the model construction are as follows:(1)Different regional characteristics will have a certain impact on people's travel behavior.Only a single model will produce certain deviations.(2)Research shows that the IIA characteristics will cause certain deviations to the construction model.Choosing the appropriate disaggregate model plays an important role in analyzing the travel behavior of urban residents.(3)Data non-response often occurs in surveys of residents travel behaviors,especially sensitive personal information.Simply deleting the missing data will cause a large deviation from the model establishment and needs to be corrected by the data imputation method.(4)Under the influence of multiple differences(regional differences,model differences and data differences),it is also the biggest contribution of this paper to explore the key influencing factors of urban residents travel behavior and its law of action.