Natural Gas Price Forecasting by a New Hybrid Model Combining Quadratic Decomposition Technology and LSTM Model

Abstract

Research on the price prediction of natural gas is of great significance to market participants of all kinds. In order to predict natural gas prices more reliably, this paper introduces a quadratic decomposition technology based on the combination of variational modal decomposition (VMD) and ensemble empirical modal decomposition (EEMD), which decomposes the residual term (Res) after VMD by EEMD; then, a new hybrid model called VMD-EEMD-Res.-LSTM is constructed in combination with the long short-term memory (LSTM) prediction model. The contribution of this new hybrid model is that, unlike existing application research that combines existing decomposition technology with the LSTM model, it does not ignore the important information contained in the residual after the VMD. In order to verify the predictive performance of the proposed new model, this paper uses the data of the spot price of natural gas in the United States to conduct a multistep-ahead empirical comparative analysis. The results show that the new hybrid model constructed in this paper has significant predictive advantages.

Advancing Natural Gas Price Predictions with ConcaveLSTM

“…Through a strategic integration of Stacked LSTM's depth and Bidirectional LSTM's bidirectional learning capabilities, the ConcaveLSTM model achieved significant improvements in predictive accuracy. Compared to previous methodologies, such as the hybrid model combining quadratic decomposition with LSTM networks proposed by Zhan and Tang [7], which aimed to tackle the non-linearity in natural gas price data, the ConcaveLSTM model demonstrated superior performance. Specifically, configurations utilizing 50 input steps paired with either 100 or 300 neurons stood out, offering the lowest prediction errors and highest congruence with real market movements.…”

“…Studies by Yang and Choi [14], leveraging machine learning algorithms for forecasting spot LNG prices, and Chen et al [5], exploring the unpredictability of natural gas prices amidst uncertainties, exemplify this trend. The effectiveness of combining computational techniques was further illustrated by Zhan and Tang [7] through their hybrid model, which underscores the field's progression toward more accurate and comprehensive forecasting methods.…”

“…This early study set the stage for further exploration into sophisticated predictive models, aiming to enhance accuracy and reliability. Following this, Zhan and Tang [7] introduced a hybrid model that combines quadratic decomposition with Long Short-Term Memory (LSTM) networks, specifically addressing the challenge of nonlinearity in natural gas price data. This approach represents a significant advancement in the field, demonstrating the potential of combining statistical techniques with machine learning to improve forecasting accuracy.…”

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

“…Results indicate that the model had the highest prediction results and outperformed previous research. Zhan et al (2022) proposed a technique combining LSTM and quadratic decomposition to construct an improved natural gas forecasting hybrid model VMD-EEMD-Res.-LSTM and then selected the monthly natural gas spot price of Henry Hub for empirical analysis. Wang et al (2022) combined the decomposition algorithm with the multi-objective grasshopper optimization algorithm, and used nine models to predict the decomposed components.…”

A Study on Agricultural Commodity Price Prediction Model Based on Secondary Decomposition and Long Short-Term Memory Network

“…To further enhance predictive accuracy, a complex system theory approach that combines decomposition techniques with forecasting models has been proposed by Wang Shouyang [24]. This approach can effectively overcome the limitations of traditional econometric models and artificial intelligence models [25]. In fact, in most combination forecasting models, signal processing methods are used to decompose time series, and intelligent models are employed to predict the components after decomposition [26].…”