The evolution of import in Turkey and an analysis abaut predictability of import by artificial neural network method

In this study, purpose is to examine historical evolution of imports is Turkey and to determine the best architecture to predict import by Artificial Neural Network method. In this context, structure of imports and its relation with other economic variables are examined and an analysis about the predictability of imports by Artificial Neural Network method is conducted. Artificial Neural Network method is used in this study, which is not widely used in social sciences in Turkey. Explanatory variables include Expenditure based Gross Domestic Product, CPI based, Real Effective Foreign Exchange Rate, and Export for goods and service, Banking Sector Credit Volume. Quarterly data is used for the period of 1980-2007. Realized data and forecast data have been compered, the best performance was determined architecture. Later in the year 1980-2012 with data import data was estimated for the years 2013-2014-2015. The obtained results, in the estimation of import of Artificial Neural Networks Method of modeling techniques have shown that good.