Improving the prediction accuracy in electric load forecasting is an important goal to be pursued in order to optimize the management of economic and environmental resources. We propose in this paper a customized prediction approach, which relies on the chaotic behavior of the electric load time series and on the spectral characteristics of its prediction error. The proposed predictor is based on a twofold prediction scheme using Mixture of Gaussian neural networks.
Improving accuracy of electric load short-term forecasting by using MoG neural networks / Panella, Massimo; FRATTALE MASCIOLI, Fabio Massimo; Rizzi, Antonello; Martinelli, Giuseppe. - In: ATTI DELLA FONDAZIONE GIORGIO RONCHI. - ISSN 0391-2051. - STAMPA. - LVII (4):(2002), pp. 689-692. (Intervento presentato al convegno Applicazioni delle Reti Neurali nell’Ingegneria Elettrica ed Elettromagnetica (Giornata di Studio) tenutosi a Firenze nel 4-5 aprile 2002).
Improving accuracy of electric load short-term forecasting by using MoG neural networks
PANELLA, Massimo;FRATTALE MASCIOLI, Fabio Massimo;RIZZI, Antonello;MARTINELLI, Giuseppe
2002
Abstract
Improving the prediction accuracy in electric load forecasting is an important goal to be pursued in order to optimize the management of economic and environmental resources. We propose in this paper a customized prediction approach, which relies on the chaotic behavior of the electric load time series and on the spectral characteristics of its prediction error. The proposed predictor is based on a twofold prediction scheme using Mixture of Gaussian neural networks.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
