عنوان مقاله [English]
Prediction is one of the most important achievements of modeling science, which has a special place in management and decision making. In general, there is a direct relationship between the accuracy of predictions and the quality of made decisions. This is the most important reason why efforts for providing more precise methods of prediction in the subject literature have not stopped despite the existence of numerous methods. The classical Auto-Regressive Integrated Moving Average (ARIMA) models are one of the most important and well-known statistical methods that have been frequently used in various sciences. However, these methods, despite all their unique advantages, have some disadvantages, which sometimes reduce their acceptability. One of the most important of these disadvantages is the limitation of the linearity, the limitation of certainty, the limitation of the number of required data, and the limitation of mixed and multiple structures. Many attempts have been made to address these shortcomings and limitations in the literature. In this paper, a method for overcoming the limitation of complex and multiple structures is presented using the Ensemble Empirical Mode Decomposition (EEMD) techniques. In the proposed method, at first, the under-study time series, which is essentially complex and involves several simultaneous structures, is decomposed into its constituent constituents, which are fundamentally less complicated and include fewer structures. Then, each of these simplified structures is
predicted using an auto-regressive integrated moving average model. Ultimately, the prediction of each of the main components is combined to formulate final predictions. The results of applying the proposed method to predict the crude oil price, which is among the most complex time series in financial markets, indicate the effectiveness of the proposed method. Numerical results show that the proposed method can improve the performance of the classic auto-regressive integrated moving average of 65.57% and 53.85% in predicting Texas and Brent crude oil prices.