Predição Multiescala de Tráfego de Redes Utilizando Redes Neurais RBF Treinadas com Algoritmo de Mínimos Quadrados Ortogonais
DOI:
https://doi.org/10.5540/tema.2008.09.03.0503Abstract
Neste artigo, apresentamos um novo algoritmo de treinamento para redes neurais RBF (Função de Base Radial) baseado em mínimos quadrados ortogonais e em decomposição multiescala de sinais. Propomos um algoritmo de predição de séries temporais que combina as predições de aproximações para estas mesmas séries e de seus detalhes em diferentes escalas através da Transformada Wavelet. Aplicamos as redes neurais RBF treinadas com o algoritmo proposto na prediçãode tráfego de redes de computadores. O treinamento das redes neurais RBF com algoritmo de mínimos quadrados ortogonais aliada à decomposição wavelet contribui para evitar problemas de mal condicionamento da matriz de interpolação, como também para melhorar a capacidade de extrapolação da rede neural RBF. Esta última característica é verificada pela redução do erro quadrático médio de predição. As simulações realizadas confirmam que predições mais precisas são obtidas para as séries temporais de tráfego de redes em relação a outras redes neurais existentes.References
S. Chen, S.A. Billings, Neural networks for nonlinear dynamic system modeling and identification, International Journal of Control 56, No. 2 (1992), 319-346.
G. Cybenco, Approximations by superposition of a sigmoidal function, Math. Control Signal Systems 2, No. 4 (1989), 303-314.
S. Haykin, “Neural Networks - A Comprehensive Foundation”, Prentice Hall, 2ed., 1998.
X. He, A. Lapedes, “Nonlinear Modeling and Prediction by Successive Approximations using Radial Basis Functions”, Technical Report, Los Alamos National Laboratory, 1991.
J.-S.R. Jamg, C.-T. Sun, Functional equivalence between radial basis function networks and fuzzy inference systems, IEEE Transactions on Neural Networks, 4 (1993), 156-159.
W.W.Y. Ng, A. Dorado, D.S. Yeung, W. Pedrycz, E. Izquierdo, Image classification with the use of radial basis function neural networks and the minimization of the localized generalization error. Pattern Recognition 40, No. 1 (2007), 19-32.
M.J.L. Orr, Regularization in the selection of radial basis function centres, Neural Computation, 7, No. 3 (1995), 606-623.
M.A.S. Potts, D.S. Broomhead, Time series prediciotn with a radial basis function neural network, SPIE Adaptive Signal Processing, 1565 (1991), 255-266.
M.J.D. Powell. Radial basis function approximations to polynomials, in “Proc. 12th Hiennial Numerical Analysis Conf.” (Dundee), pp. 223-241, 1987.
A. Sang, S.Q. Li, A predictability analysis of network traffic. in “Conference on Computer Communications”, IEEE Infocom, New York, Mar. 2000.
D. Veitch, P. Abry, A wavelet based joint estimator of the parameters of longrange dependence, IEEE Trans. Inform. Theory–Special Issue on Multiscale Statistical Signal Analysis and Its Applications 45, No. 3 (1999), 878-897.
F.H.T. Vieira, “Predição de Tráfego em Redes de Comunicações utilizando Redes Neurais e Análise Wavelet - Alocação Dinâmica de Largura de Faixa”. Dissertação de mestrado. Universidade Federal de Goiás, Goiânia, Goiás, Brasil,2002.
D.F. Walnut, “An Introduction to Wavelet Analysis”. Birkh¨auser Boston, 1ed., 2004.
Downloads
Published
How to Cite
Issue
Section
License
Authors who publish in this journal agree to the following terms:
Authors retain copyright and grant the journal the right of first publication, with the work simultaneously licensed under the Creative Commons Attribution License that allows the sharing of the work with acknowledgment of authorship and initial publication in this journal.
Authors are authorized to assume additional contracts separately, for non-exclusive distribution of the version of the work published in this journal (eg, publish in an institutional repository or as a book chapter), with acknowledgment of authorship and initial publication in this journal.
Authors are allowed and encouraged to publish and distribute their work online (eg, in institutional repositories or on their personal page) at any point before or during the editorial process, as this can generate productive changes as well as increase impact and the citation of the published work (See The effect of open access).
This is an open access journal which means that all content is freely available without charge to the user or his/her institution. Users are allowed to read, download, copy, distribute, print, search, or link to the full texts of the articles, or use them for any other lawful purpose, without asking prior permission from the publisher or the
author. This is in accordance with the BOAI definition of open access
Intellectual Property
All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License under attribution BY.