The Rough Set Approach to Association Rule Mining
J. W. Guan,D. A. Bell,D. Y. Liu
DOI: https://doi.org/10.1109/ICDM.2003.1250969
2003-01-01
Abstract:In transaction processing, an association is said to existbetween two sets of items when a transaction containingone set is likely to also contain the other. In informationretrieval, an association between two sets of keywords occurswhen they co-occur in a document. Similarly, in datamining, an association occurs when one attribute set occurstogether with another. As the number of such associationsmay be large, maximal association rules are sought, e.g.,Feldman et al (1997, 1998).Rough set theory is a successful tool for data mining. Byusing this theory, rules similar to maximal associations canbe found. However, we show that the rough set approach todiscovering knowledge is much simpler than the maximalassociation method.