Analysis and Prediction of Spatiotemporal Distribution of Accidents at Highway-rail Grade Crossings Caused by Different Causes

To study the spatiotemporal distribution differences of highway-rail grade crossing (HRGC) accidents caused by different causes and provide data support for the prevention of HRGC accidents. HRGC accidents were first subdivided into 6 categories according to accident causes in this study. Then, based on O-test and C-test, the characteristics of different types of HRGC accident data are tested to select appropriate models. Finally, suitable generalized linear regression models were developed to predict the frequency of HRGC accidents caused by different causes in different spatiotemporal conditions. The results show that the constructed prediction models can accurately predict the spatiotemporal distribution of HRGC accidents caused by different causes. The research can provide data support for the formulation of accident prevention strategies and the allocation of improved funds of HRGC.