Please use this identifier to cite or link to this item:
Title: Art networks with geometrical distances
Authors: Dagher, Issam 
Affiliations: Department of Computer Engineering 
Keywords: L-p
Fuzzy ART
Category choice
Subjects: Neural networks (Computer science)
Issue Date: 2006
Part of: Journal of discrete algorithms
Volume: 4
Issue: 4
Start page: 538
End page: 553
In this paper, ART networks (Fuzzy ART and Fuzzy ARTMAP) with geometrical norms are presented. The category choice of these networks is based on the L-p norm. Geometrical properties of these architectures are presented. Comparisons between this category choice and the category choice of the ART networks are illustrated. And simulation results on the databases taken from the UCI repository are performed. It will be shown that using the L-p norm is geometrically more attractive. It will operate directly on the input patterns without the need for doing any preprocessing. It should be noted that the ART architecture requires two preprocessing steps: normalization and complement coding. Simulation results on different databases show the good generalization performance of the Fuzzy ARTMAP with L-p norm compared to the performance of a typical Fuzzy ARTMAP.
This paper was presented in " The IASTED International Conference on Artificial Intelligence and Applications (AIA 2005). February 14 to February 16, 2005. Innsbruck, Austria. ".
DOI: 10.1016/j.jda.2005.06.007
Ezproxy URL: Link to full text
Type: Journal Article
Appears in Collections:Department of Computer Engineering

Show full item record


checked on Oct 23, 2021

Record view(s)

checked on Oct 23, 2021

Google ScholarTM


Dimensions Altmetric

Dimensions Altmetric

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.