Henry Hub monthly natural gas price forecasting using CEEMDAN-Bagging-HHO-SVR

As a clean fossil energy source, natural gas plays a crucial role in the global energy transition. Forecasting natural gas prices is an important area of research. This paper aims at developing a novel hybrid model that contributes to the prediction of natural gas prices. We develop a novel hybrid model that combines the "Decomposition Algorithm" (CEEMDAN), "Ensemble Algorithm" (Bagging), "Optimization Algorithm" (HHO), and "Forecasting model" (SVR). The hybrid model is used for monthly Henry Hub natural gas forecasting. To avoid the problem of data leakage caused by decomposing the whole time series, we propose a rolling decomposition algorithm. In addition, we analyzed the factors affecting Henry Hub natural gas prices for multivariate forecasting. Experimental results indicate that the proposed model is more effective than the traditional model at predicting natural gas prices.