Prediction Crude Oil Supply in the Eleven Producing Countries: Use of Neural Networks and Linear Regression (1980-2006)

اقتصاد پولی مالی

مقاله 5، دوره 16، شماره 27، آبان 1388   

نوع مقاله: پژوهشی

شناسه دیجیتال (DOI): 10.22067/pm.v16i27.1990

نویسندگان

شکیبایی شکیبایی^* ؛ نظام آبادی پور نظام آبادی پور؛ حسینی حسینی

چکیده

Most oligopolistic models in the oil market begin with the assumption that OPEC is a cartel. Successive increases in oil prices during recent years, and OPEC’s inability to regularize the oil market fortifies the possibility of the existence of competitive behavior in the oil market. Now, with regard to the intense fluctuations in the oil market, prediction of oil supply for politicians and oil companies is important. In the recent years, use of neural networks, along with econometrical methods, of forecasting of economic variables is a norm. Because neural networks have, exclusive characteristics like release of statistical hypotheses and capability of solving complex nonlinear problems. In this article in addition to prediction oil supply in the eleven producing countries, by use of linear regression and neural networks, the results of every country are separately compared. The estimation results show that neural networks are presenting better predictions in comparison to the linear regression models.