Multiobjective Genetic Algorithm for Extracting Subgroup Discovery Fuzzy Rules

Author

del Jesus, María José ; González, Pedro ; Herrera, Francisco

Author_Institution

Dept. of Comput. Sci., Jaen Univ.

fYear

2007

fDate

1-5 April 2007

Firstpage

50

Lastpage

57

Abstract

This paper presents a multiobjective genetic algorithm for obtaining fuzzy rules for subgroup discovery. This kind of fuzzy rules lets us represent knowledge about patterns of interest in an explanatory and understandable form which can be used by the expert. The multiobjective algorithm proposed in this paper defines three objectives. One of them is used as a restriction on the rules in order to obtain a Pareto front composed of a set of quite different rules with a high degree of coverage over the examples. The other two objectives take into account the support and the confidence of the rules. The use of the mentioned objective as restriction allows us the extraction of a set of rules which describe more complete information on most of the examples. Experimental evaluation of the algorithm, applying it to a market problem shows the validity of the proposal obtaining novel and valuable knowledge for the experts

Keywords

data mining; fuzzy set theory; genetic algorithms; Pareto front; multiobjective genetic algorithm; subgroup discovery fuzzy rule extraction; Bibliographies; Computational intelligence; Computer science; Data mining; Decision making; Fuzzy reasoning; Fuzzy sets; Fuzzy systems; Genetic algorithms; Proposals;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Intelligence in Multicriteria Decision Making, IEEE Symposium on

Conference_Location

Honolulu, HI

Print_ISBN

1-4244-0702-8

Type

conf

DOI

10.1109/MCDM.2007.369416

Filename

4222982