Integrating classification and association rule mining-英文文献.pdf

Appeared in KDD-98, New York, Aug 27-31, 1998 Integrating Classification and Association Rule Mining Bing Liu Wynne Hsu Yiming Ma Department of Information Systems and Computer Science National University of Singapore Lower Kent Ridge Road, Singapore 119260 {liub, whsu, mayiming }@.sg Abstract rules as the class association rules (CARs). An existing Classification rule mining aims to discover a small set of association rule mining algorithm (Agrawal and Srikant rules in the database that forms an accurate classifier. 1994) is adapted to mine all the CARs that satisfy the Association rule mining finds all the rules existing in the minimum support and minimum confidence constraints. database that satisfy some minimum support and minimum This adaptation is necessary for two main reasons: confidence constraints. For association rule mining, the 1. Unlike a transactional database normally used in target of discovery is not pre-determined, while for association rule mining (Agrawal and Srikant 1994) that classification rule mining there is one and only one pre- does not have many associations, classification data determined target. In this paper, we propose to