Title :
Reconstructed dynamics and chaotic signal modeling
Author :
Kuo, Jyh-Ming ; Principe, Jose C.
Author_Institution :
Comput. NeuroEng. Lab., Florida Univ., Gainesville, FL, USA
Abstract :
A nonlinear AR model is derived from the reconstructed dynamics of a signal. The underlying system is assumed to be nonlinear, autonomous, and deterministic. In this formulation, the output error scheme is shown to be more suitable than the equation error scheme in training a network as a nonlinear AR model of the signal. A method to incorporate the information of the dynamical invariants in signal modeling is proposed
Keywords :
autoregressive processes; chaos; learning (artificial intelligence); neural nets; nonlinear systems; signal processing; state-space methods; chaotic signal modeling; dynamical invariants; network training sequence; neural network; nonlinear AR model; output error; reconstructed dynamics; state space trajectory; Chaos; Delay; Frequency domain analysis; History; Length measurement; Multilayer perceptrons; Nonlinear dynamical systems; Nonlinear equations; Predictive models; System testing;
Conference_Titel :
Neural Networks for Signal Processing [1994] IV. Proceedings of the 1994 IEEE Workshop
Conference_Location :
Ermioni
Print_ISBN :
0-7803-2026-3
DOI :
10.1109/NNSP.1994.365999
