Practical estimation of traffic accidents at intersections by neural network.

The safety project has been considered in terms of cost-effectiveness with budget constraints. The basic concept is to choose the proper intersection improved as a profitable traffic safety management. The accurate estimation for traffic accident with several factors should be proposed. The neural network (NN) is applied to describe the non-linear relationship between traffic accidents and the physical factors. The numerical example is given to show the performance of the approach. As the practical estimation model is created, the impact of traffic safety project can be measured though the change of traffic conditions at intersections. For the covering abstract see ITRD E114174.