Analysis of driver injury severity in truck-involved accidents using a non-parametric classification tree model

摘要: To explore the factors contributing to driver injury severity in traffic accidents, parametric regression models, such as multinomial logit models (MNLs) or ordered probabilistic have been commonly applied for many years. However, these their own model assumptions and pre-defined underlying relationships between dependent independent variables. If are violated, can lead erroneous estimation of likelihood. This study collects 2005–2006 truck-involved accident data from national freeways Taiwan develops a non-parametric Classification Regression Tree (CART) establish empirical relationship outcomes driver/vehicle characteristics, highway geometric variables, environmental The results show that drinking-driving, seatbelt use, vehicle type, collision circumstance action, number vehicles involved location were key determinants truck accidents.