Using the neural networks approach to measure the impact of transport variables on economic development in Syria

Abstract

Transportation is essential in improving the market economy and revitalizing the economy in all its sectors. Therefore, this research aims to propose a neural network model to determine the impact of transport variables in all its sectors on economic development in Syria, and to compare transport efficiency in Syria at the level of variables and sectors. A multilayer perceptron (MLP) network is used that reduces prediction error of target (also called output) variables. The results showed that the variable of vehicles per km of roads is the most important with a rate of 100% and an effect of 0.188, that is, every change of one unit in this variable leads to a change of 0.188 units in the indicators of economic and social development, then we find in the second place the variable of the number of passengers per kilometer for rail transport At a rate of 86.3%, with an impact score of 0.163 for each unit of economic and social development variables. Then comes the variable of the number of ships that arrived at the Syrian ports, the third most important variable, at a rate of 43.1%, with an impact score of 0.081 for each unit of the economic and social development variables, and the variable of the number of rail transport modes comes in the fourth rank, with a rate of 41% and an impact score of 0.077, and in the fifth and sixth rank, respectively, variables Air transport: The number of passengers and the number of aircraft landing, by 37.6% and 35.7%, respectively. The table shows that the variables of land and rail transport are more efficient than sea and air transport.