The number of vehicle has been a well-known traffic variable that is used in the estimation of number of deaths in traffic accidents all along. But due to the insufficiency of the number of vehicles, it cannot build a strong relationship just as vehicle-mile does. In this study, Smeed equation is improved by the usage of vehicle-mile instead of vehicle number. Then forecasting dead number models were developed using Artificial Neural Network (ANN) method which has the ability to model complex nonlinear relationships between variables without a prior assumption. Different ANN models were constructed to obtain the best ANN form by changing the transfer functions, train algorithms, and neuron numbers. At the end of the study, it was observed that the correlation coefficients and the mean square error values of the best regression and ANN models increased from 0.936 to 0.970 and decreased from 41 960 to 20 407, respectively. It can be concluded that the forecasting ability is improved using the vehicle-mile variable on the logarithmic regression and the ANN models of death number in traffic accidents.

Login Required

Please login to access the full PDF of this article.

🔑 Author Login