Using Neural Networks in Time Series to Forecast Oil Prices in Iraq

Abstract

This research represents an attempt by the researchers to use the new mechanism not traditional in the prediction of oil prices in Iraq, using neural networks, it is clear, that any model or any formula or any data can be expressed in a time series can be used in the characterization of the subject of expectations through All previous observations or variables counter weight chronologically. under the enormous development in the software field Where the simulation of the work of the neuron was reached where called neural network, where it is identifying the problem under study and then an occasion of neural network design work, and that the neural network work is based on the education and training of the network to reduce the error coefficient to the lowest level by adjusting the value of the weights midwife to estimate in advance the model known structure, where they were in this study the use of artificial neural networks to predict the price of oil for five years to come is (2016,2017, 2018,2019, 2020) where it was to get these prices, the research found that the use of neural networks is better than normal methods , so as to overcome the basic assumptions in the estimation.