抄録
The rear-end crash on highways tends to continuously increase. This is in accordance with many researches pointing out that the collisions are likely to highly occur at the intersection as vehicles have to reduce the velocity to approach into the intersection. Therefore, this research focuses on seeking for the ways decreasing the number of rear-end collisions at the intersection through the guidelines for relevant organizations to improve such as physical features of roads as well as the promotion methods. The accident cases occurring on highways from 2011 to 2015 were the data used to be analyzed by classification and regression tree (CRT). The target variables were rear-end crashes at the intersection/and those outside the intersection. The results from tree model found that the significant variables to be further recommended were: Average traffic volume, road surface type and lighting condition factor. In addition, this research has provided the guidelines for reducing the number of crashes at the intersection. It also guides the study of rear-end crashes at the intersections in the future.
