Forecasting Realized Volatility of Chinese Crude Oil Futures with a New Secondary Decomposition Ensemble Learning Approach

This study proposes a new VMD-ICEEMDAN-LSTM model, that combines secondary decomposition with long short-term memory neural networks (LSTM) to forecast realized volatility (RV) of Chinese crude oil futures. First, RV sequence is decomposed into sub-components and residuals using variational mode decomposition (VMD). Then, the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) is applied to perform a secondary decomposition on the residuals. Finally, LSTM was used to forecast all the decomposed components, and then combine all the forecasting results to obtain our final forecasting value. Our results show that VMD-ICEEMDAN-LSTM model significantly outperforms existing individual and combination models