Title: Prediction of the bottled propane gas sales using a neural network-based model

Authors: Horacio Paggi; Franco Robledo; Aldo Portela

Addresses: School of Advanced Computer Engineering (Escuela Superior de Ingenieros Informáticos), Madrid Politechnical University (Universidad Politécnica de Madrid), Madrid, Spain ' IMERL-INCO, Engineering Faculty, University of the Republic (Universidad de la República), Montevideo, Uruguay ' IMERL, Engineering Faculty, University of the Republic (Universidad de la República), Montevideo, Uruguay

Abstract: This work presents an application of the artificial neural networks (ANN) in the prediction of the time series of the weekly wholesale of bottled propane gas (13 kg bottles). For this purpose, several networks with different topologies were built. Many schemas of ensembles were applied to reduce prediction of errors. Additionally, given the scarce data available, it was mandatory to minimise the input dimensionality of the networks and to do this, with a rational and systematic approach, considerations about stochastic dynamical systems were made and generalisations of the Takens-Mañé's theorem for non-deterministic systems were used.

Keywords: stochastic dynamical systems; Takens-Mañé's theorem application; time series prediction; bottled propane gas sales; artificial neural networks; ANNs; modelling; non-deterministic systems.

DOI: 10.1504/IJMHEUR.2016.081155

International Journal of Metaheuristics, 2016 Vol.5 No.3/4, pp.254 - 277

Received: 29 Dec 2015
Accepted: 07 Jul 2016

Published online: 24 Dec 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article