Recherche de concepts à partir de données arborescentes et imprécises

摘要

n this article, we propose a formalism (ASN) to deal with imprecise and structured data described with attributes and imprecise values. The ASN allow us to represente entities that are composed with parts and sub-parts ; values may be imprecise, unknown and the attributes may be not applicable. We can also take into account constraints that exist between the values of the attributes. We aim to find concepts from a set of entities described with ASN. Concepts are defined from an extension of the Galois lattice theory to deal with imprecise and structured data. To find concepts, we propose an incremental algorithm that compute a lattice concepts extracted from the Galois lattice where the too general concepts? in regard to a given criteria? are not computed.