Search

Urban-rural integration is the critical process of urbanization, while during this process there are various corresponding problems generated, safety is the most important one, especially at urban-rural fringe areas due to mixed traffic flow and complicated land use. This study intended to investigate the injury severity interaction between urban area and urban-rural fringe area, and explore the correlation within injury severity levels and heterogeneity attributed to unobserved factors. To address the correlation and heterogeneity issues, a grouped random parameters seemingly unrelated bivariate (SUB) probit model was proposed, in which the SUB probit model addressed the correlation of residuals, while the random parameters model accommodated the heterogeneity due to unobserved factors. By comparing the pooled with random parameters models, the results showed that random parameters SUB probit model performed better than the pooled one, and the dataset collected from 2013 to 2017 in Chengdu, China was adopted to illustrate the proposed model. It is found that crash location, speed limit and age of person injured are significant for injury severity in urban and urban-rural fringe areas, but crash form plays a significant role in urban area while number of persons involved should be paid more attention due to injury severity in urban-rural fringe area. Some empirical suggestions are presented to improve the safety in urban and urban-rural fringe areas.

With the development of big data technology and the improvement of deep learning technology, data-driven and machine learning application have been widely employed. By adopting the data-driven machine learning method, with the help of clustering processing of data sets, a recurrent neural network (RNN) model based on Keras framework is proposed to predict the injury severity in urban areas. First, with crash data from 2014 to 2017 in Nevada, OPTICS clustering algorithm is employed to extract the crash injury in Las Vegas. Next, by virtue of Keras' high efficiency and strong scalability, the parameters of loss function, activation function and optimizer of the deep learning model are determined to realize the training of the model and the visualization of the training results, and the RNN model is constructed. Finally, on the basis of training and testing data, the model can predict the injury severity with high accuracy and high training speed. The results provide an alternative and some potential insights on the injury severity prediction.