par Akritas, Pavlos;Akishin, Pavel P.G.;Antoniou, Ioannis 
;Bonushkina, Yu;Drossinos, Yannis I.;Ivanov, Victor V.V.;Kalinovsky, Yu Y.L.;Korenkov, Vladimir V.V.;Zrelov, P.V.
Référence Chaos, solitons and fractals, 17, 2-3, page (305-309)
Publication Publié, 2003-07
;Bonushkina, Yu;Drossinos, Yannis I.;Ivanov, Victor V.V.;Kalinovsky, Yu Y.L.;Korenkov, Vladimir V.V.;Zrelov, P.V.Référence Chaos, solitons and fractals, 17, 2-3, page (305-309)
Publication Publié, 2003-07
Article révisé par les pairs
| Résumé : | We applied a nonlinear time series analysis to the traffic measurements, obtained at the input of a medium size local area network. In order to reconstruct the underlying dynamical system, we estimated the correlation length τ and the embedding dimension dE of the traffic time series. In order to extract the regular part from traffic data, we filtered out the high frequency, "noisy" part. The reliable values of τ and dE permitted to apply a layered neural network for the identification and reconstruction of the underlying dynamical system. © 2002 Elsevier Science Ltd. All rights reserved. |
