Research on Travelers' Daily Activity-Travel Behavior Based on Competing Risk Model

摘要

The study of urban residents' time allocation behaviors of daily activity-travel helps to understand their decision-making mechanisms, and take corresponding measures to effectively guide travelers' travel behavior. Based on the analysis of travel data of ATUS in the U.S., this paper extracted the influencing factors of travelers' activity scheduling. The joint selection process of activity participation type and activity duration on the continuous time axis is explained through sensitivity analysis. The results show that: (1) According to the activity scheduling behavior of travelers, it can be found that each trip has its next trip associated with it based on activity chain. (2) The covariate parameter estimation results based on Cox model show that factors such as age, race, marital status, family income, family size, activity date, total number of daily activities, and duration time of activity N+l have significant influence on the duration time of activity N. (3) The estimated model suggesting that there is interrelated and constraint relationship between activity N and activity N+l of travelers. Moreover, the strength of that relations changes significantly with type of activities. The competing risk models used in this paper are a useful tool for describing such differences. These fundamental works are help for enriching and expanding the travel behavior analysis theory and method.