Study On Students’ Commute Mode Choice Behavior Based On Logit Model And Empirical Sensitivity

As an important part of urban commute,studentse commute has brought tremendous pressure on urban traffic because of its strong travel regularity.In the complicated travel environment,students are still immature in physical and mental development,resulting in frequent traffic accidents with students as the mainstay.At the same time,facing the decline of students’ physical fitness,most people believe that optimizing the students’ commute structure and promoting active transportation can promote students’ physical activity and health.Therefore,the study of students’commute mode choice has important reference significance for alleviating urban congestion,improving the travel environment and promoting students’ active transportation.In this paper,students’ commute mode choice behavior is taken as the research object,MNL model and NL model are compared to analyze,and the empirical sensitivity based on the elastic analysis is introduced to evaluate the prediction results.Firstly,on the basis of the existing research,the paper determines the influencing factors of students5 commute mode choice from three aspects:personal,family and travel.And divide walking,cycling,public transportation and car into active transportation and other transportation.Taking the students’ commute data from Anshan residents’ travel survey as the sample,and conducting descriptive statistical analysis and correlation statistical analysis.Then establishing MNL model and NL model of students9 comnute mode choice,using R language to solve the model.And the models were compared and evaluated by five indexes of LL’AIC’BIC,AICc and R2.Secondly,the students’commute mode choice model with two interactive items of gender and travel distance,income and travel distance is established,and explained the influence of univariate and interactive variables by using the empirical sensitivity.Finally,the corresponding management strategies are proposed according to the research results.The study found that the five indexes of NL model are better than MNL model,so NL model has better statistical performance,and empirical sensitivity can be used to evaluate any number of changes in variables.For different grades of students,there are differences in the commute mode choice,especially for the choice of cycling presents a 4tdumbbell’ structure.When the travel distance is increased by 1km and 2km,the corresponding empirical sensitivity is 31.9%and 51.4%,which indicates that a small increase in the distance can promote the use of bicycles.Public transportation and car commuting have more competitive advantages when traveling long distances.There is an obvious interaction between income and travel distance.With the increase of distance,the increase of income will accelerate the decline choice of walking and promote the choice of cars,but the students in high-income families do not conform to this rule.These findings could have implications for promoting active transportation for students and serve as references for traffic planning and management.