Road Network Safety Ranking Using Accident Prediction Models

Road network safety ranking procedure and its implementation is not a strictly regulated road safety activity. This is one of the most flexible ways to determine the most effective and beneficious road safety investments. This paper analyses possibilities for implementing road network safety ranking according to the accidents prediction. The application of road network safety ranking procedure enables to prevent road accidents, i.e. to implement a proactive road safety activity. Road accident predictions are made using the empirical Bayes method, which is based on the assumption that in a similar environment with the similar traffic conditions the risk of accidents is similar. In order to implement this method, all roads of national significance of Lithuania were divided into homogeneous road sections and junctions. The homogenous road groups were determined based on 2012‒2016 data of geometrical parameters of the road and traffic volume. Having estimated the predicted number of accidents for each homogenous road section, it is possible to calculate the predicted accident rate for each road. The authors of the paper, have predicted accident rate for the whole road, compiled a map of road safety levels for the trans-European roads in Lithuanian.