Forecasting Crude Oil Prices using a Hybrid Model Combining Long Short-Term Memory Neural Networks and Markov Switching Model

Given the significant impact of crude oil prices on the global economy, accurately predicting their fluctuations is essential for effective decision-making in the energy sector. Therefore, this research aims to develop a hybrid model that can comprehensively capture the nonlinear and volatile characteristics of crude oil prices and provide accurate predictions. The proposed approach involves segmenting the time series into multiple sub-series, which capture the nonlinear and volatile characteristics of crude oil prices. The nonlinear sub-series is predicted using Long Short-Term Memory neural networks, while the volatile and fluctuating sub-series are forecasted using a Markov Switching model. The results of these predictions are combined using a linear combination to estimate the crude oil price time series. The proposed hybrid model provides a comprehensive understanding of the various factors that drive crude oil price fluctuations, making it a valuable tool for decision-making in the energy sector.