Un nouveau modèle pour la classification non supervisée sous contraintes

Constrained clustering is an important task in Data Mining. In the last ten years, many works have been done to extend classical clustering algorithms to handle user-defined constraints, but they are in general limited to one kind of constraints. In our previous work (Dao et al., 2013a), we have proposed a declarative and general framework, based on Constraint Programming, which enables to design a clustering task by specifying an optimization criterion and different kinds of user-constraints. The model is designed for a clustering task, where data is divided in exactly k clusters. In this paper, we present a new model for constrained clustering tasks where the number of clusters is only bounded. It offers more flexibility, while being more efficient and experiments show that it has a better performance, when compared to existing complete methods dealing with the same criteria. MOTS-CLÉS : classification non supervisée, contraintes utilisateur, programmation par contraintes.