Title :
A comparative study of the category choice of the fuzzy ART with the L-1 norm
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Balamand, Lebanon
Abstract :
In this paper, a comparative study of the category choice of the Fuzzy ART with the L-1 norm is presented. It is shown that the category choice can be replaced by a distance measure related to the L-1 norm. This distance measure will have the following advantages over the category choice of the Fuzzy ART network: 1) no need for augmenting the dimensions of the input patterns. The distance measure will operate directly on the input patterns without the need for doing complement coding; and 2) no need for normalizing the input patterns. The input patterns need not to be in the interval.
Keywords :
ART neural nets; category theory; fuzzy neural nets; learning (artificial intelligence); adaptive resonant theory; category choice; complement coding; distance measure; fuzzy ART networks; normalization; Clustering algorithms; Data preprocessing; Equations; Subspace constraints;
Conference_Titel :
Neural Networks, 2003. Proceedings of the International Joint Conference on
Print_ISBN :
0-7803-7898-9
DOI :
10.1109/IJCNN.2003.1223709
