Comparing The Econometric Methods and Artificial Neural Networks in Predict of Corn Import of Iran

The importance of predicting the economic variables for policy makers and program planners is no secret to anyone. Therefore, over recent decades, several competing models have been developed. In this study, the amount of corn import to Iran is predicted for the period 2010-2014 by using econometric techniques and artificial neural networks methods. The results show that the feed forward neural network with fewer errors and better performance compared with econometric techniques ARIMA and exponential smoothing to predict the amount of corn imports. The study results also show that the amount of corn imports would grow in 2011 by 8 percent less than that of the previous year. Also the highest reduction in corn imports compared with that of the previous year by the amount of 11 percent decrease is related to the year 2012. So, it is necessary to increase domestic production to meet domestic needs.

JEL Classification: D12, C32, C22

Keywords:
Predict, Neural Network, Corn, Import, Iran