پيش بيني پوياي قيمت نفت خام با استفاده از شبكه هاي عصبي مصنوعي و با به كارگيري ذخيره سازي هاي نفتي كشورهاي OECD

Forecasting of Crude Oil Pricjes Using Neural Networks and OECD Inventories

On one hand, oil is the greatest energy resource in the world and, on the other hand, because of the role of oil revenue in the economic of oil producer countries, such as Iran,it is vital for these countries. So it is necessary to recognize different affective parameters on oil market for these countries. In this research, we try to forecast oil price as an important variable in world wide oil market by using neural networks and ARIMA model. The results of dynamic forecasts have shown that in all cases neural network has better results than ARIMA model. In addition, the results of this research have shown that by use of OECD inventories as an added input in model and doing a bivariate forecasting (for the first time in Iran) the error of oil prices forecasts will reduce. JEL Classification: C02, Q40, Q41, C22, C45