Detecting oscillations in neural networks via frequency domain analysis
Jorge L. Moiola, DANIEL W. BERNS, Guanrong Chen, Haluk Öğmen · Proceedings of International Conference on Neural Networks (ICNN'97) · 2002
We offer a graphical analysis for detecting orbital oscillations in a neural network model. The approach taken is based on a frequency domain methodology, in which graphical techniques are used to capture some of the intrinsic dynamical features of periodic solutions in a nonlinear neural network system with time delays. The graphical analysis provides a correct detection of large-amplitude and multiple limit cycles existing in the network model.