Research On Intercity Travel Behavior Based On Big Data And Latent Class-Mixed Logit Choice Model

Intercity transportation corridor is an important undertaker to realize the exchange of people and materials between cities.With the continuous advancement of regional development and integration,the demand for passenger and freight in intercity transportation corridors is increasing day by day.Improving the transportation structure is one of the important means to improve the transportation capacity of transportation corridors,which relies on accurate research conclusions on intercity travel behavior.Mobile phone signaling data has the advantages of low collection cost,large sample size,and wide coverage,and it has great application potential in travel behavior research.By comparing and analyzing the characteristics of questionnaire data and cell phone signaling data,combined with domestic and foreign precedents using cell phone signaling data to solve traffic problems,a series of cell phone signaling data processing methods have been formulated,including intercity travel user screening and trajectory chain extraction.,Recognition of stop points and other preprocessing procedures,as well as the process of extracting traffic information such as traveler’s residence,intercity travel mode,travel time,and travel cost.On the basis of the existing data set,the latent category model is used to classify and cluster big data,which realizes the potential classification of intercity travelers and supplements the personal attributes of intercity travelers from the group level.A mixed logit model of inter-city travel mode choice behavior was constructed to analyze the travel characteristics of different potential groups.Choices include private cars,intercity buses,regular-speed railways,and high-speed railways.The explanatory variables include gender,age,living location,travel time,travel costs,travel efficiency,weekly average number of intercity trips,and the highest number of destination cities(Second)There are nine long stay points and stay time.During the demonstration,based on the mobile phone signaling data of Sichuan Province from November 4 to November 10,2019,a total of 64512 intercity travelers from Chengdu to Mianyang were extracted.The potential category model is used to classify the population,and two potential categories are obtained: "low-age,high-consumption group" and "higher-age,low-consumption group".It is found that the "low-age,high-consumption group" has more trips per week,and this group is more inclined Choose private cars to travel,while the "older age group with low consumption" is more inclined to choose high-speed rail.The mixed logit model was used to explain the inter-city travel mode choice behaviors of the two potential categories of people,and it was found that the influencing factors of the "low-age and high-consumption group" and the "middle-age and low-consumption group" were different.The “low-age and high-consumption group” inter-city travelers are greatly affected by age,living location,average weekly number of inter-city trips,longest stay at the point of stay,travel efficiency,inter-city travel time and cost.Inter-city travelers of the "older age and low consumption group" are greatly affected by gender,age,residential location,the nature of the longest stay at the destination,average weekly intercity travel time,travel expenses and travel time.