| With the development of social economy,the travel demand of urban residents is growing rapidly.As a new type of travel mode,online car-hailing has experienced several years of development,which has a certain degree of impact on the travel structure of residents.The introduction of online car hailing,a new type of travel mode,alleviates the difficulty of taxi hailing and provides residents with diversified and multi-level travel mode choices.It makes up for the shortcomings of the traditional mode of transportation,improves the travel conditions of residents,but also affects the urban traffic network structure to a certain extent.In the face of the scarcity of road resources,we should coordinate the development of car-hailing and other modes of transportation to meet the different travel needs of residents.Therefore,this paper needs to study and analyze the different travel behaviors of residents.Firstly,according to the travel survey data of resident online car-hailing users in 2019,the individual attributes,travel characteristics and travel mode characteristics of residents are statistically analyzed;through the difference analysis of residents’ travel characteristics,it is found that the vehicle ownership in family attributes,gender,age,education background,occupation and personal income in personal attributes,and the frequency of express use,the number of trips and the severity in travel characteristics According to the time requirement,the travel time,travel cost,discount and premium in travel mode characteristics are closely related to the travel mode choice of Dalian residents.Secondly,NL model and support vector machine model are constructed.NL model is mainly used to study the factors that affect the travel mode choice of online car-hailing users in Dalian.The results show that personal attributes,travel characteristics and travel mode characteristics all affect the travel mode choice of online car-hailing users in Dalian.Based on the results of NL1 model,this paper analyzes the sensitivity of travel mode selection based on elasticity,and analyzes the impact on travel mode from four aspects: reducing travel time,reducing travel cost,increasing discount and reducing premium.Then,build support vector machine model.The support vector machine model divides the samples into training set and test set,obtains a group of parameters with the highest prediction accuracy through training,and forecasts the test set,and the total prediction accuracy is more than 95%.Compared with NL model,the prediction accuracy of SVM model is higher.Finally,according to the factors of travel time,travel cost,discount and premium,this paper puts forward the travel optimization and improvement scheme for Dalian residents’ car-hailing users.The support vector machine(SVM)model is used to evaluate the effect of travel optimization improvement scheme for car-hailing users in Dalian residential network. |
