DocumentCode :
303277
Title :
A new fast U-D factorization-based learning algorithm with applications to nonlinear system modeling and identification
Author :
Zhang, Youmin ; Li, X. Rong
Author_Institution :
Dept. of Electr. Eng., New Orleans Univ., LA, USA
Volume :
1
fYear :
1996
fDate :
3-6 Jun 1996
Firstpage :
623
Abstract :
A new fast learning algorithm for training multilayer feedforward neural networks by using variable forgetting factor technique and U-D factorization-based fading memory extended Kalman filter is proposed. In comparison with the backpropagation (BP) and extended Kalman filter (EKF) based learning algorithms, the proposed algorithm can provide much more accurate learning results in fewer iterations with fewer hidden nodes as well as improve convergence rate and numerical stability. In, addition, it is less sensitive to the choice of initial weights and initial covariance matrix as well as other setup parameters. Simulation results of nonlinear dynamic system modeling and identification show that the new algorithm proposed here is an effective and efficient learning algorithm for feedforward neural networks
Keywords :
Kalman filters; feedforward neural nets; filtering theory; identification; learning (artificial intelligence); modelling; multilayer perceptrons; nonlinear systems; U-D factorization-based fading memory extended Kalman filter; U-D factorization-based learning algorithm; convergence rate; feedforward neural networks; initial covariance matrix; initial weights; multilayer feedforward neural networks training; nonlinear dynamic system; nonlinear system identification; nonlinear system modeling; numerical stability; variable forgetting factor technique; Backpropagation algorithms; Convergence of numerical methods; Covariance matrix; Fading; Feedforward neural networks; Modeling; Multi-layer neural network; Neural networks; Nonlinear dynamical systems; Numerical stability;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Neural Networks, 1996., IEEE International Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-3210-5
Type :
conf
DOI :
10.1109/ICNN.1996.548967
Filename :
548967
Link To Document :
بازگشت