|
Title:
|
Acceleration of the gradient-type methods in blind source separation |
|
Author(s):
|
David Blanco , Diego P. Ruiz , María C. Carrion , Carlos García-puntonet |
|
ISBN:
|
ISSN: 1646-3692 |
|
Editors:
|
Mohammad Essaaidi and Mohammed El Mohajir (Guest Editors) |
|
Year:
|
2009 |
|
Edition:
|
V IV, 1 |
|
Keywords:
|
radar target recognition, natural resonances, principal component analysis, neural network, matrix pencil method. |
|
Type:
|
Journal Paper |
|
First Page:
|
21 |
|
Last Page:
|
29 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
The steepest descendent is the non-linear optimiza-tion method most used in ICA algorithms. The method is used to find the unmixing matrix, which solves the problem and is a minimum of a non-linear cost function. In this paper the use of quasi-Newton optimization methods, instead of gradient-type ICA methods, is studied. These methods can increase the speed of the method what it is corroborated by simulations. |
|
|
|
|
|
|
