Estimating the revenues of istanbul bosphorus bridges

Province of Istanbul is Turkey's most populous city. With the expansion of the city, the need for a transition between the two sides increases. Depending on the increasing needs, 3 bridges were built in the city of Istanbul. Depending on the number of vehicles passing through these bridges, estimating revenues is of great importance in planning future transportation investments. In this study, income prediction models are created depending on the vehicles passing through the July 15 Martyrs Bridge and Fatih Sultan Mehmet Bridge in the Bosphorus. The analysis methods used in this study are artificial neural networks (ANN) and multivariate linear regression (MLR) methods. n these analyzes, annual income is calculated based on the variables of population, dollar rate, net minimum wage, total expense and total number of vehicles crossing the bridge. When the results of the linear method of the MLR model are analyzed, the correlation coefficient (R) is 0.9746, mean squares of errors (MSE) value is 603047344915727 and mean percent of errors (MPE) is calculated as 8.51%. Considering the results of the Interaction method, R value is calculated as 1.00, MSE value is 1.2*10-11 and MPE value is calculated as 4*10-12%. Looking at the results of the Purequadratic method, R value is calculated as 0.9956, MSE value is 106006178076566 and MPE value is calculated as 3.52%. As a result of the analysis made with ANN, the model with 14 neurons, tansig transfer function, Levenberg-Marquardt training algorithm gives the best results. According to these results, the R value obtained is calculated as 0.9649, MPE value is 9.6665% and the value of MSE is calculated as 923528331783739. According to these results, the method that gives the best results depending on R, MSE and MPE criteria is multivariate linear regression method.