Influence and conditional influence-new interestingness measures in association rule mining

Author

Chen, Guoqing ; Liu, De ; Li, Jiexun

Author_Institution

Sch. of Econ. & Manage., Tsinghua Univ., Beijing, China

Volume

3

fYear

2001

fDate

6/23/1905 12:00:00 AM

Firstpage

1440

Lastpage

1443

Abstract

Discusses the issues of interestingness in association rule mining. First, a rule is possibly redundant or misleading even if it possesses high degrees of confidence and support. Second, association rules do not reflect the effect of negatively influential facts. Such problems are related to confidence deviation. In the paper, therefore, two new measures of interestingness, namely influence and conditional influence, are introduced to represent the effect of the antecedent on the consequent. Furthermore, the mining algorithms are extended accordingly such that certain redundant rules can be eliminated and negatively influential rules may be discovered

Keywords

data mining; fuzzy logic; fuzzy set theory; antecedent; association rule mining; conditional influence; confidence deviation; consequent; degrees of confidence; degrees of support; interestingness measures; negatively influential facts; Association rules; Data mining; Decision making; Filtering; Itemsets; Transaction databases;

fLanguage

English

Publisher

ieee

Conference_Titel

Fuzzy Systems, 2001. The 10th IEEE International Conference on

Conference_Location

Melbourne, Vic.

Print_ISBN

0-7803-7293-X

Type

conf

DOI

10.1109/FUZZ.2001.1008930

Filename

1008930