Artificial Intelligence Review

, Volume 16, Issue 3, pp 177–199

Understanding the Crucial Role of Attribute Interaction in Data Mining

  • Alex A. Freitas

DOI: 10.1023/A:1011996210207

Cite this article as:
Freitas, A.A. Artificial Intelligence Review (2001) 16: 177. doi:10.1023/A:1011996210207


This is a review paper, whose goal is tosignificantly improve our understanding of thecrucial role of attribute interaction in datamining. The main contributions of this paperare as follows. Firstly, we show that theconcept of attribute interaction has a crucialrole across different kinds of problem in datamining, such as attribute construction, copingwith small disjuncts, induction of first-orderlogic rules, detection of Simpson's paradox,and finding several types of interesting rules.Hence, a better understanding of attributeinteraction can lead to a better understandingof the relationship between these kinds ofproblems, which are usually studied separatelyfrom each other. Secondly, we draw attention tothe fact that most rule induction algorithmsare based on a greedy search which does notcope well with the problem of attributeinteraction, and point out some alternativekinds of rule discovery methods which tend tocope better with this problem. Thirdly, wediscussed several algorithms and methods fordiscovering interesting knowledge that,implicitly or explicitly, are based on theconcept of attribute interaction.

attribute interaction classification constructive induction data mining evolutionary algorithms inductive logic programming rule induction rule interestingness Simpson's paradox small disjuncts 

Copyright information

© Kluwer Academic Publishers 2001

Authors and Affiliations

  • Alex A. Freitas
    • 1
  1. 1.Postgraduate Program in Computer SciencePontificia Universidade Catolica do ParanaCuritiba, PRBrazil (E-mail

Personalised recommendations