Saturated systems of homogeneous boxes and the logical analysis of numerical data

Authors:
P. L. Hammer;Y. Liu;B. Simeone;S. Szedmák
Affiliations:
RUTCOR, Rutgers Center of Operations Research, Piscataway, NJ;School of Science, University of Ontario Institute of Technology, 2000 Simcoe Street North, Oshawa, Ont, Canada L1H 7K4;Department of Statistics, "La Sapienza" University, Piazzale Aldo Moro 5, 00185, Rome, Italy;Electronics and Computer Science, University of Southampton, UK
Venue:
Discrete Applied Mathematics - Discrete mathematics & data mining (DM & DM)
Year:
2004

Citing 3
Cited 2

Cause-effect relationships and partially defined Boolean functions

Annals of Operations Research
Machine Learning

Machine Learning
An Implementation of Logical Analysis of Data

IEEE Transactions on Knowledge and Data Engineering

Analysis of a multi-category classifier

Discrete Applied Mathematics
Classifying negative and positive points by optimal box clustering

Discrete Applied Mathematics

Quantified Score

Hi-index	0.00

Visualization

Abstract

Following the general principles of the logical analysis of data methodology, originally developed for the case of binary data, we define a similar approach for the analysis of numerical data. The central concepts of this methodology are those of homogeneous boxes and of saturated systems of homogeneous boxes. The box-clustering heuristic described in this paper is efficient and was applied successfully for the analysis of datasets concerning breast tumors, oil exploration and diabetes.