Article ID Journal Published Year Pages File Type
4948409 Neurocomputing 2016 6 Pages PDF
Abstract
This letter investigates the stability analysis of complex ICA by Negentropy Maximization. Based on the theoretical analysis for the fixed-point iteration and cost function of Quasi-Newton CMN (N-CMN) algorithm, we have found that (1) the unstable area does not have any influence on the performance of the N-CMN algorithm and (2) the algorithm may converge to an undesirable solution even if the sources satisfy the stable conditions.
Keywords
Related Topics
Physical Sciences and Engineering Computer Science Artificial Intelligence
Authors
, ,