X. He and H. Asada, "A new method for identifying orders of inputoutput models for nonlinear dynamic system," in Proceedings of the American Control Conference, San Francisco, California USA, 1993, pp. 2520--2523.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....et l indice SBF 250. 1. INTRODUCTION La prdiction d une srie temporelle est un problme trs courant, que ce soit dans le domaine industriel (prdiction de la consommation lectrique d un pays, dbit d une rivire. ou dans le domaine financier (prdiction de taux de changes, cours boursier, [2,5,8,9]. Malheureusement, la prdiction doit tre effectue en l absence d indications sur le processus sous jacent ces sries temporelles. D 38 A. Lendasse, E. De Bodt, A. Choppin, M. Verleysen Association Connectioniste en Sciences Economiques et de Gestion, ACSEG 98 Louvain la Neuve (Belgium) ....

....pas. Il y a deux faons d interprter ces dfinitions. La premire, qui est la plus courante, consiste considrer que les lments du vecteur autor gressif sont des lments bruts qui proviennent directement de la srie temporelle. Il faut donc rechercher quels sont ces lments et leur nombre optimal [5,9]. Une seconde interprtation de ces dfinitions consiste supposer qu il existe un vecteur auto rgressif idal (c est dire le plus petit possible) mais que celui ci est en fait compos de combinaisons (non linaires) de valeurs de la srie brute, et non des valeurs elles mmes. Dans la suite, nous ....

Xiangdong He and Haruhiko Asada, "A new method for Identifying Orders of InputOutput Models for Nonlinear Dynamic Systems," Proc. of the American Control Conf., S.F., California, 1993 pp. 2520-2523.

....the group A data. combined with a search tree to find the inputs of the model. None of the described process is very efficient because the parameters of the model has to be calculated every time a candidate regressor is tested. A good alternative is a model free test proposed by He and Asada in [11]. The method is based in the evaluation of the so called Lipschitz Quotients. A Lipschitz quotient is defined in this framework: given a nonlinear function y = f(x) and N input output pairs (y i ; x) i ) the Lipschitz quotient q ij is given by, q ij = jy i Gamma y j j jx i Gamma x j j ; i ....

X. He and H. Asada, " A New Method for Identifying Orders of Input-Output Models for Nonlinear Dynamic Systems ,"Proceedings of the American Control Conference pp.2520-2523, San Francisco California, June 1993

....the embedding space of a time series. In the neural networks literature, Pi and Peterson [29] have introduced the ffi test to address this issue. In system identification, the use of Lipschitz quotients has been proposed in order to identify the order of non linear input output systems [15]. A similar method was applied to time series, under the name of geometrical technique , in the signal processing literature [24] Though different in practice, these methods rely on a common assumption on the continuity and relative smoothness of the underlying mapping g, and use a ....

X. He and H. Asada. A new method for identifying orders of input-output models for nonlinear dynamic systems. In American Conference on Control, San Francisco, California, 1993.

....[16] which is essentially equivalent to finding the set of primary delays in time series. In the realm of neural computation, the recently proposed ffi test method[14] addresses this issue. In a different field, a method for identifying the order of non linear input output systems was proposed [5], that relies on the use of Lipschitz quotients i.e. ratio between output and input distances. A similar method applied to time series (called geometrical technique ) was presented last year at this workshop [10] Though different in practice, these methods rely on a common assumption on the ....

X. He and H. Asada, "A new method for identifying orders of inputoutput models for nonlinear dynamic systems," in American Conference on Control, San Francisco, California, 1993.

No context found.

X. He and H. Asada, "A new method for identifying orders of inputoutput models for nonlinear dynamic system," in Proceedings of the American Control Conference, San Francisco, California USA, 1993, pp. 2520--2523.

No context found.

X. He and H. Asada, A New Method for Identifying Orders of Input-Output Models for Nonlinear Dynamic Systems(Proc. ACC pp.2520, S.Fco Cal.1993)

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC