In this paper, we propose to investigate the notion of integrity constraints in inductive databases. We advocate that integrity constraints can be used in this context as an abstract concept to encompass common data mining tasks such as the detection of corrupted data or of patterns that contradict the expert beliefs. To illustrate this possibility we propose a form of constraints called association map constraints to specify authorized confidence variations among the association rules. These constraints are easy to read and thus can be used to write clear specifications. We also present experiments showing that their satisfaction can be tested in practice. © Springer-Verlag Berlin Heidelberg 2004.
CITATION STYLE
Bykowski, A., Daurel, T., Méger, N., & Rigotti, C. (2004). Integrity constraints over association rules. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2682, 306–323. https://doi.org/10.1007/978-3-540-44497-8_16
Mendeley helps you to discover research relevant for your work.