Title :
A recurrent Newton algorithm and its convergence properties
Author :
Kuan, Chung-Ming
Author_Institution :
Dept. of Econ., Nat. Taiwan Univ., Taipei, Taiwan
fDate :
5/1/1995 12:00:00 AM
Abstract :
In this paper a recurrent Newton algorithm for an important class of recurrent neural networks is introduced. It is noted that a suitable constraint must be imposed on recurrent variables to ensure proper convergence behavior. The simulation results show that the proposed Newton algorithm with the suggested constraint performs uniformly better than the backpropagation algorithm and the Newton algorithm without the constraint, in terms of mean-squared errors
Keywords :
Newton method; convergence of numerical methods; learning (artificial intelligence); recurrent neural nets; convergence properties; mean-squared errors; recurrent Newton algorithm; recurrent neural networks; Backpropagation algorithms; Convergence; Feedforward neural networks; Neural networks; Output feedback; Process control; Recurrent neural networks; Signal processing algorithms; System identification; Target recognition;
Journal_Title :
Neural Networks, IEEE Transactions on
