Title :
Application of Prony signal analysis to recurrent neural networks
Author :
M. Farrokhi;C. Isik
Author_Institution :
Dept. of Electr. & Comput. Eng., Syracuse Univ., NY, USA
Abstract :
Recurrent neural networks are highly nonlinear dynamic systems, and therefore it is not an easy task to analyze their dynamic behavior. The primary goal of this paper is to apply Prony signal analysis and a modified root locus technique to recurrent neural networks. The Prony method is compared with other linearization methods to analyze recurrent neural networks, and their performances are demonstrated using simulated examples.
Keywords :
"Signal analysis","Recurrent neural networks","Poles and zeros","Stability","Nonlinear dynamical systems","Equations","Nonlinear systems","Eigenvalues and eigenfunctions","Application software","Computer networks"
Conference_Titel :
Neural Networks, 1994. IEEE World Congress on Computational Intelligence., 1994 IEEE International Conference on
Print_ISBN :
0-7803-1901-X
DOI :
10.1109/ICNN.1994.374198
