Elmira Emsia - Academic Staff of Accounting Department, Department of Accounting, Damghan Branch, Islamic Azad University, Damghan, Iran.

Gross Domestic Production (GDP) plays significant role in the economic growth. Hence, understanding the future status of this factor invigorate us to make better decision for the futurein term of economic policy. In the recent years, Artificial Neural Networks (ANNs) have shown great ability in prediction of economic-based time series. This paper seeks acomparison between neural network and econometric approaches to predict GDP growth in different countries (developing and developed countries) based on time series datacollected from 1990– 2012. To this end, two intelligence-based scenarios named Artificial Neural Networks (ANN) and self organizing map (SOM) method are developed in which theparameters of the ANN model are to be optimized through. The obtained results showed that the ANN is superior to other techniques and models such as OLS in prediction of GDP.

GDP growth, ANN, trade, developed and developing countries