Density of Closed Balls in Real-Valued and Autometrized Boolean Spaces for Clustering Applications

Authors:
C. G. González;W. Bonventi, Jr.;A. L. Rodrigues
Affiliations:
Technology Center, University of Sorocaba, Brazil;Technology Center, University of Sorocaba, Brazil;Technology Center, University of Sorocaba, Brazil
Venue:
SBIA '08 Proceedings of the 19th Brazilian Symposium on Artificial Intelligence: Advances in Artificial Intelligence
Year:
2008

Citing 5
Cited 0

Data clustering: a review

ACM Computing Surveys (CSUR)
ROCK: a robust clustering algorithm for categorical attributes

Information Systems
Clustering and its validation in a symbolic framework

Pattern Recognition Letters
Topology for Computing (Cambridge Monographs on Applied and Computational Mathematics)

Topology for Computing (Cambridge Monographs on Applied and Computational Mathematics)
Pattern Recognition, Third Edition

Pattern Recognition, Third Edition

Quantified Score

Hi-index	0.00

Visualization

Abstract

The use of real-valued distances between bit vectors is customary in clustering applications. However, there is another, rarely used, kind of distances on bit vector spaces: the autometrized Boolean-valued distances, taking values in the same Boolean algebra, instead of 茂戮驴. In this paper we use the topological concept of closed ball to define density in regions of the bit vector space and then introduce two algorithms to compare these different sorts of distances. A few, initial experiments using public databases, are consistent with the hypothesis that Boolean distances can yield a better classification, but more experiments are necessary to confirm it.