Performance evaluation of different neural network training algorithms in error control coding

摘要

This paper discusses the error control coding in communication networks using a neural network approach. A comparison was made between two different learning algorithms - namely, back propagation and random optimization - for this application. The results indicate the superiority of back propagation over random optimization for learning time and generalization. It has been found that the error control coding problem is an auto-associative problem and is best suited for back propagation environments.