Liquid Neural Networks Improve Natural Gas Price Forecasting

A recent study published on arXiv investigates the use of Liquid Neural Networks (LNNs) for short-term predictions of the Henry Hub natural gas spot price. These LNNs are designed to continuously adjust to changing temporal patterns via dynamic updates of their internal states, making them well-suited for the fluctuating and nonstationary nature of natural gas prices, affected by seasonal demand, geopolitical factors, and macroeconomic changes. The goal of this research is to improve forecasting precision in unstable markets, ultimately minimizing uncertainty in energy trading and power market applications. This paper falls under the category of Computer Science > Machine Learning.