Preferentially ordered clustering

Patrick Meyer (1, 2) , Alexandru Liviu Olteanu (1, 3)

1 Lab-STICC_TB_CID_DECIDE
2 LUSSI - Département Logique des Usages, Sciences sociales et Sciences de l'Information
3 University of Luxembourg [Luxembourg]

Clustering is defined as the process of grouping objects that are similar and separating those that are dissimilar. In the field of Multi-Criteria Decision Aid (MCDA) these objects are defined in a richer con-text and compared with respect to the preferences of a Decision Maker(DM). Clustering is closely related to the classical MCDA problems of ranking and sorting. As a result only a few attempts have been made so far at exploring this problem in detail. We model in this article the problem of preferentially ordered clustering using an objective function and propose a method that looks for the optimal result with respect to it. We diversify this objective through an additional property which could be desirable.

Mots clés: Clustering; Multicriteria decision aiding

Domaines: Recherche opérationnelle [math.OC]

Patrick Meyer, Alexandru Liviu Olteanu. Preferentially ordered clustering. MDAI 2012: 9th Modeling Decisions for Artificial Intelligence Conference, Nov 2012, Girona, Spain. pp.87 - 98. ⟨hal-00800945⟩