posted on 2023-05-26, 16:37authored byMirkazemi, Aram
This report presents an investigation on the convergence behaviour of Backpropagation and Kohonen neural networks. The causes of convergence problems have been identified and current techniques for overcoming them are reported. Two new techniques are also presented for improvement of the convergence rate of the Backpropagation neural network. The outcome of this investigation is encapsulated in a simulation software package which includes most of the current techniques and those proposed by the author.
History
Publication status
Unpublished
Rights statement
Copyright 1990 the Author - The University is continuing to endeavour to trace the copyright owner(s) and in the meantime this item has been reproduced here in good faith. We would be pleased to hear from the copyright owner(s). Thesis (MEngSc(Qual))--University of Tasmania, 1990. Includes bibliographical references