Proceedings of the 21st Benelux Conference on Artificial Intelligence pages:315-316
Benelux Conference on Artificial Intelligence edition:21 location:Eindhoven, the Netherlands date:29-30 October 2009
Discovering itemsets and conjunctive rules under constraints are popular topics in the data mining and
machine learning communities, for which many algorithms have been proposed. Despite the abundance of
research in this area, however, constraint programming (CP) techniques developed in the artificial intelli-
gence community to deal with constraint satisfaction problems have never been applied to rule discovery.
In , we show that CP can not only be applied in an intuitive, extendible way to rule discovery, but also
that CP techniques significantly outperform existing approaches in data mining.