On meteorological forecasts for energy management and large historical data: A first look

Abstract

This communication is devoted to a com-parison between various meteorological forecasts, for the pur-pose of energy management, via different time series tech-niques. The first group of methods necessitates a large num-ber of historical data. The second one does not and is mucheasier to implement, although its performances are todayonly slightly inferior. Theoretical justifications are related tomethods stemming from a new approach to time series, artifi-cial neural networks, computational intelligence and machinelearning. Several numerical simulations are provided and dis-cussed.