Application of Long Short-Term Memory Networks for Natural Gas Demand Prediction

Natural gas consumption has seen unprecedented increment worldwide due to its reliability and economic behavior. Limited gas reserves lead to precise energy planning and distribution. A state’s stratum depends on its energy supply and devising for a certain demand. A forecasting model for natural gas consumption can predict gas demand and make the policymakers manage the strategy for the future. This work proposed a neural network-based model to estimate the natural gas demand for upcoming months using a Long Short Term Memory network. The socio-economic variables are considered input for forecasting monthly natural gas consumption. Evaluation of the approach is based on the real-life dataset. The study’s findings signify that the Long Short Term Memory model is reliable and accurate, with lower errors for time consumption projection.