Nouvel article accepté dans un journal indersceince: "International Journal of Modelling, Identification and Control"

Meteorological time series forecasting with pruned multi-layer perceptron and 2-stage Levenberg-Marquardt method

cyril voyant, Wani Tamas, Marie-Laure Nivet, Gilles Notton, Christophe Paoli, Aurélia Balu, Marc Muselli

Cet article fait suite au travail mené et valorisé lors de l'ICM² 2014 de Prague. La nouvelle méthodologie de pruning exposée semble donner de bon résultats, il convient maintenant de la comparer avec les méthodes existantes afin de d'établir une hiérarchisation des algorithmes sur le cas réel de séries temporelles de rayonnement global. L'abstract :

"A Multi-Layer Perceptron (MLP) defines a family of artificial neural networks often used in TS modeling and forecasting. Because of its “black box” aspect, many researchers refuse to use it. Moreover, the optimization (often based on the exhaustive approach where “all” configurations are tested) and learning phases of this artificial intelligence tool (often based on the Levenberg-Marquardt algorithm; LMA) are weaknesses of this approach (exhaustively and local minima). These two tasks must be repeated depending on the knowledge of each new problem studied, making the process, long, laborious and not systematically robust. In this paper a pruning process is proposed. This method allows, during the training phase, to carry out an inputs selecting method activating (or not) inter-nodes connections in order to verify if forecasting is improved. We propose to use iteratively the popular damped least-squares method to activate inputs and neurons. A first pass is applied to 10% of the learning sample to determine weights significantly different from 0 and delete other. Then a classical batch process based on LMA is used with the new MLP. The validation is done using 25 measured meteorological TS and cross-comparing the prediction results of the classical LMA and the 2-stage LMA."

Rédigé par Cyril VOYANT le Mardi 8 Juillet 2014 à 15:29 | Lu 399 fois