Prediction of energy analysis parameters in transportation systems

Turkey's energy consumption values aren't separating freight and passenger. Therefore, there are freight and passenger separation coefficients by making use of the energy consumption values of Canada. This adds energy consumption values based on the type of Turkey transportation sector are allocated as freight and passenger and energy density value can be calculated. When the energy density values in the result of these accounts examined Turkey seems to be a growing inefficiencies in the transport sector. Estimations of energy consumption values based on the modes of transportation sector in Turkey are carried out using the methods of Logarithmic Regression, Multivariate Linear Regression and Artificial Neural Networks (ANN),. The model that gives the best results from the methods used in the analyzes for highways is the quadratic regression type which is one of the multivariate linear regression types. For quadratic regression model, the mean square of errors (MSE) value is 5806018987,9117, the mean percentage error (MPE) value is 0,5068% and the correlation coefficient (R) value is 0,9999. The model that gives the best results from the methods used in the analyzes for the railways is the ANN model. For ANN model, MSE value is 63870745,2578, MPE value is 2,2466% and R value is 0,9064. ANN method is the method that gives the best results for the analyzes used for maritime. As a result of ANN analysis, MSE value is 1786726958,4754, MPE value is 4,6616% and R value is 0,9637. For the airlines, the ANN method is the method that gives the best results for the estimation analysis. As a result of ANN analysis, the MSE value was 26341153522,5271, the MPE value was 10,6019% and the R value was 0,9417. Among the statistical and artificial intelligence models created as a result of these analyzes, the quadratic regression model on the highways gives the best results, while the ANN model on the railways, maritime and airlines gives the best results.