Long-Term Energy Load Forecasting Using Auto-Regressive and Approximating Support Vector Regression

Several techniques can be retrieved in literature, which cope with the problem of energy load forecasting in the short-term framework, that is hours up to few days ahead. However, in order to properly schedule operative conditions including e- nergy purchasing and generation, fuel supply, and infrastruc- ture development and maintenance, a long-term prediction is of crucial importance. In this framework, the generaliza- tion of the short-term techniques to this more complex prob- lem is usually characterized by a scarce performance. In this paper, we present an innovative method, which exploits the Savitzky-Golay filter and the Support Vector Regression al- gorithm in order to reliably predict the energy consumption in the long-term framework: the extrapolation ability of our proposal is validated on a real-world problem.