Rough Sets Based Rule Generation from Data with Categorical and Numerical Values

Rough set theory has been mainly applied to data with categorical values. In order to handle data with numerical values in this theory, a familiar concept of ‘wildcards’ was employed, and a new framework of rough sets based rule generation has been proposed. Two characters @ and # were introduced in...

Full description

Saved in:
Bibliographic Details
Published in:Journal of advanced computational intelligence and intelligent informatics 2008-09, Vol.12 (5), p.426-434
Main Authors: Sakai, Hiroshi, Koba, Kazuhiro, Nakata, Michinori
Format: Article
Language:English
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Rough set theory has been mainly applied to data with categorical values. In order to handle data with numerical values in this theory, a familiar concept of ‘wildcards’ was employed, and a new framework of rough sets based rule generation has been proposed. Two characters @ and # were introduced into this framework, and numerical patterns were also defined for numerical values. The concepts of ‘coarse’ and ‘fine’ for rules were explicitly defined according to numerical patterns. This paper enhances the previous framework, and describes the implementation of an utility program. This utility program is applied to the data in UCI Machine Learning Repository, and some useful rules are obtained.
ISSN:1343-0130
1883-8014
DOI:10.20965/jaciii.2008.p0426