Conférence EGC 2009 (Extraction et Gestion des Connaissances), Strasbourg, France, 27-30 janvier 2009
Abstract
In this paper, Formal Concept Analysis and Galois lattices are used for the analysis of complex datasets, online social networks in particular. Lattice-inspired statistics computed on the objects of the lattice provide their "conceptual distribution". An experimentation conducted on four social networks' samples shows how these statistics may be used to characterize these networks and filter them automatically.