On recurrent neural networks for auto-similar traffic prediction: a performance evaluation

Menezes, José Wally Mendonça de; Barreto, Guilherme de Alencar

Use este identificador para citar ou linkar para este item: http://repositorio.ufc.br/handle/riufc/70695

Registro completo de metadados

Campo DC	Valor	Idioma
dc.contributor.author	Menezes, José Wally Mendonça de	-
dc.contributor.author	Barreto, Guilherme de Alencar	-
dc.date.accessioned	2023-02-09T16:14:17Z	-
dc.date.available	2023-02-09T16:14:17Z	-
dc.date.issued	2006	-
dc.identifier.citation	MENEZES, J. W. M.; BARRETO, G. A. On recurrent neural networks for auto-similar traffic prediction: a performance evaluation. In: INTERNATIONAL TELECOMMUNICATIONS SYMPOSIUM, 2006, Fortaleza. Anais... Fortaleza: IEEE, 2006. p. 534-539.	pt_BR
dc.identifier.uri	http://www.repositorio.ufc.br/handle/riufc/70695	-
dc.language.iso	en	pt_BR
dc.publisher	International Telecommunications Symposium	pt_BR
dc.subject	Recurrent neural networks	pt_BR
dc.subject	Traffic prediction	pt_BR
dc.subject	Auto-similar processes	pt_BR
dc.subject	VBR video traffic	pt_BR
dc.subject	Multi-step-ahead prediction	pt_BR
dc.title	On recurrent neural networks for auto-similar traffic prediction: a performance evaluation	pt_BR
dc.type	Artigo de Evento	pt_BR
dc.description.abstract-ptbr	The NARX network is a recurrent neural architecture commonly used for input-output modelling of nonlinear systems. The input of the NARX network is formed by two tapped-delay lines, one sliding over the input signal and the other one over the output signal. Currently, when applied to nonlinear time series prediction, the NARX architecture is designed as a plain Focused Time Delay Neural Network (FTDNN); thus, limiting its predictive abilities. In this paper, we propose a strategy that allows the original NARX architecture to fully exploit its computational resources to improve prediction performance. We use real-world VBR video traffic time series to evaluate the proposed approach in multi-step-ahead prediction tasks. The results show that the proposed approach consistently outperforms standard neural network based predictors, such as the FTDNN and Elman architectures.	pt_BR
Aparece nas coleções:	DETE - Trabalhos apresentados em eventos

Arquivos associados a este item:

Arquivo	Descrição	Tamanho	Formato
2006_eve_gabarreto.pdf		416,63 kB	Adobe PDF	Visualizar/Abrir

Mostrar registro simples do item Visualizar estatísticas