Research On Commuter's Departure Time Based On Survival Analysis Approach

The travel of commuter is not only a high proportion of urban residents' travel behavior,but also is the most basic and important behaviors in individual life.The commuting travel during early morning will make the road traffic congestion worse.The departure time of commuting behavior can reflect the commuting traffic itself regularity.Also,it is an important parameter for temporal and spatial distribution.So it is necessary to do research on the departure time.This paper will focus on the department time and analyze it from an individual point of view.Paper uses questionnaires to get the data of commuter travel behavior,based on the the theory of survival analysis,and extract some factors that will influence the department time.Then Cox hazard-model of commuting departure time is established.According to the regression results,we ca n get specific influence of the relevant variables and use it to predict the department time.The main researchs of this paper are as follows:(1)This paper explored the overall departure time distribution of commuting.After analyzing the data of resident trip survey,it use the nonparametric method to estimates the survival function and hazard function,analyze the overall distribution of departure time.(2)This paper investigated Cox hazard-model and the influencing factors of departure time.The applicability of Cox hazard-model was discussed.Then Cox hazard-model of the department were proposed,The factors such as the gender,age,occupation,whether send children to school,travel distance,travel mode,the quality of public transportation service and the quality of individual motorized transportation service all significantly influence on departure time.(3)Based on the family characteristics,personal characteristics,traffic characteristics of commuters and the quality of transportation service perceived of individuals,the Cox hazard-model can be used to analyze if the structure of travel mode changes,how the department time will change.The research results of this paper can provide strong theoretical support for the traffic guidance and traffic control measures.