کنفرانس مهندسی برق ایران

صفحه اصلی / سی امین کنفرانس بین المللی مهندسی برق

Forecasting Crude Oil Prices using improved deep belief network (IDBN) and long-term short-term memory network (LSTM)

نویسندگان :

Mohammad Mahdi Lotfi Heravi¹ Mahsa Khorrampanah² Monireh Houshmand³

1- دانشگاه صنعتی امیرکبیر 2- دانشگاه بین‌المللی امام رضا 3- دانشگاه بین‌المللی امام رضا

کلمات کلیدی :

Crude oil price forecast،Deep learning model،Improved Deep Belief Network (IDBN)،Return Nervous Network (RNN)،Long-Term Memory Network (LSTM)

چکیده :

Historically, energy resources are of strategic importance for the economic growth and social welfare of any country. Therefore, predicting crude oil price fluctuations is an important issue. Because crude oil price changes are affected by a wide range of risk factors in crude oil markets, crude oil prices show more complex nonlinear behavior and create a higher level of risk for investors than in the past. With the popularity of the deep learning model in engineering, it has attracted significant research trends in economics and finance. In this dissertation, we propose a new method of predicting the combined price of deep-based crude oil to model nonlinear dynamics in changing the price of crude oil and predict its future change at a higher level of accuracy. The results of the experiments show that the superior performance of the model based on the proposed method against statistical article [3] is statistically significant. In general, we found that the combination of the IDBN or LSTM model lowered the MSE value to 4.65, which was 0.81% lower than the base article, indicating an improvement in prediction accuracy. Interestingly, the prediction accuracy of the basic article method is lower than the proposed combined method.