Title :
Evaluation of the feedforward neural network covariance matrix error
Author :
Abid, S. ; Fnaiech, F. ; Najim, M.
Author_Institution :
E.S.S.T.T., Tunis, Tunisia
Abstract :
This paper presents a theoretical approach for the evaluation of a feedforward neural network covariance output error matrix. It is shown how the input signals errors and the different weights errors are linked together and spread over the neural network to form the output covariance matrix error which could may be used to determine an error bound. The formulas of the output covariance matrix error is derived arising the sensitivity of the additive weight perturbations or input perturbations. The analytical formulas is validated via simulation of a function approximation example showing that the theoretical result is in agreement with simulation result
Keywords :
covariance matrices; digital simulation; error analysis; feedforward neural nets; function approximation; signal processing; additive weight perturbations; analytical formulas; computer simulation; covariance output error matrix; error bound; feedforward neural network; function approximation; input perturbations; input signal; input signals errors; multilayer neural network; perturbation model; sigmoid nonlinear element; weights errors; Analytical models; Computational modeling; Computer errors; Computer networks; Computer simulation; Covariance matrix; Feedforward neural networks; Function approximation; Multi-layer neural network; Neural networks;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2000. ICASSP '00. Proceedings. 2000 IEEE International Conference on
Conference_Location :
Istanbul
Print_ISBN :
0-7803-6293-4
DOI :
10.1109/ICASSP.2000.860151
