Natural gas demand forecasting: Development of a hybrid computational model based on artificial neural network

Recently natural gas global market attracted much attention since it is cleaner than oil and also is cheaper than renewable energy sources. However, price fluctuations, environmental concerns, technological development, unconventional resources, energy security challenges, and shipment are some of the forces made energy market more dynamic and complex in the last decade. Studying of natural gas market's demand side behavior is targeted by this research to dedicate insights about plausible trends of global natural gas demands. This paper proposed a hybrid time series model which starts with data mining based techniques to detect input features and pre-processing data, then a neural network based prediction model is used to uncover global natural gas trends. 13 different features were studied and finally 6 features were selected as the most relevant feature containing: Alternative and Nuclear Energy, CO2 Emissions, GDP per Capita, Urban Population, Natural Gas Production, Oil Consumption. Finally, the proposed prediction model overcame other competitive models refer to five different error based evaluation statistics