Evaluation Of The Effectiveness Of Neural Network Models In The Modeling Of Intra-City Highway Accidents

This paper discusses research evaluating the efficiency of neural network models in modeling inner-city highway accidents. As a case study, accident data from urban highways in Mashhad and variables related to traffic flow and road geometry are used as input variables for neural network modeling. Neural network modeling involves three steps: determining the neural network architecture, determining the transfer functions, training and errors, and creating neural network models.In this research, two neural network models were presented to estimate the number of financial and fatal accidents on inner-city highways. To evaluate the efficiency and accuracy of the models, the number of accidents estimated by the models was compared with the observed number and the value of R was used. On this basis, the presented models are suitable for estimating the number of financial and fatal accidents.