Pattern Classification Using ART 1 Network

Authors

  • Chanda Thapliyal Nautiyal, Sunita Singh, U.S. Rana

Abstract

Conventional artificial neural networks are unsuccessful in solving the plasticity- stability dilemma. For a situation where there is real network, this is exposed to regularly varying environment; it is possible that it might not get the similar training vector twice ever again [22]. Going through these circumstances, back propagation will learn nothing. In this paper pattern classification is done using ART1 network. Three different sets of seven alphabets are taken as input. The proposed algorithm successfully classifies 57.14% characters. The number of maximum clusters formed is 15 and vigilance parameter is 0.5.

Downloads

Download data is not yet available.

Downloads

Published

2020-10-30

How to Cite

Chanda Thapliyal Nautiyal, Sunita Singh, U.S. Rana. (2020). Pattern Classification Using ART 1 Network . PalArch’s Journal of Archaeology of Egypt / Egyptology, 17(9), 8038 - 8044. Retrieved from https://archives.palarch.nl/index.php/jae/article/view/5706