Optimal Time Delays in Forecasting Oil Prices Using Optimal Genetic Algorithm based Dynamic Neural Network

Oil price forecasting methods include nonlinear tools, such as Artificial Neural Network. In this study we consider the time factor in forecasting through neural networks in order to calculate the optimal delays in receiving feedback from Genetic Algorithm based Dynamic Neural Network (GADNN). We use WTI crude oil price data from 2006 to 2016 to assess forecasts produced through use of GADNN with optimal delays. We discover that genetic algorithm based dynamic neural network models that take into account the time factor, increases the accuracy of oil price forecasting compared to other existing methods through reducing the complexities of neural network design and taking into account optimized time delays.