Convergence and robustness analysis of disturbed gradient neural network for solving LMS problem

Author

Liao, Wudai ; Wang, Xingfeng ; Yang, Yuyu ; Wang, Junyan

fYear

2012

fDate

29-31 May 2012

Firstpage

398

Lastpage

401

Abstract

In this paper, we introduce a kind of method for solving least mean square problems based on the gradient neural network, including the network model construction, quantitative analysis of the network global convergence and the network convergence rate about the different activation functions. MATLAB simulation results and theoretical analysis results are accordingly consistent, which further confirm the method based on Hopfield neural network has a good effect on solving the least mean square problems.

Keywords

Hopfield neural nets; convergence; gradient methods; least mean squares methods; Hopfield neural network; LMS problem solving; MATLAB simulation; activation function; disturbed gradient neural network; least mean square problem solving; network convergence rate; network global convergence; network model construction; quantitative analysis; robustness analysis; Computational modeling; Convergence; Equations; MATLAB; Mathematical model; Neural networks; Robustness; MATLAB simulation; gradient neural network global convergence; least squares problem;

fLanguage

English

Publisher

ieee

Conference_Titel

Natural Computation (ICNC), 2012 Eighth International Conference on

Conference_Location

Chongqing

ISSN

2157-9555

Print_ISBN

978-1-4577-2130-4

Type

conf

DOI

10.1109/ICNC.2012.6234546

Filename

6234546