Final Energy Consumption Forecasting by Applying Artificial Intelligence Models

Abstract

The application of artificial neural networks has been increased in many scientific sectors the last years, with the development of new machine learning techniques and methodologies. In this research, neural networks are applied in order to build and compare neural network forecasting models for predicting the final energy consumption. Predicting the energy consumption can be very significant in public management at improving the energy management and also at designing the optimal energy planning strategies. The final energy consumption covers the energy consumption in sectors such as industry, households, transport, commerce and public management. Several architectures were examined in order to construct the optimal neural network forecasting model. The results have shown a very good prediction accuracy according to the mean squared error. The proposed methodology can provide more accurate energy consumption predictions in public and environmental decision making, and they can be used in order to help the authorities at adopting proactive measures in energy management. © 2019, Springer Nature Switzerland AG.