For road users, walking is the safest and greenest way to travel. However, with the rapid development of motorization, vehicle-pedestrian traffic accidents occur frequently. In response to this phenomenon, this paper is devoted to analyse the factors contributing to the severity of vehicle-pedestrian accidents. Based on the data of vehicle-pedestrian accidents in Chicago, USA, a dataset about 10 influencing factors is constructed. Considering the severity of accident as the classification label and considering the imbalanced distribution of samples with different labels, a 2-classification decision tree and a SVM are established to identify the severity of the accident. Then, through the model comparison, it is found that the decision tree has the best classification performance, and the overall classification accuracy can reach more than 70%. Finally, nine eigenvalues were found that had a significant impact on the severity of the accident.

CICTP