Querying Relational Concept Lattices

International audienceRelational Concept Analysis (RCA) constructs conceptual abstractions from objects described by both own properties and inter-object links, while dealing with several sorts of objects. RCA produces lattices for each category of objects and those lattices are connected via relational attributes that are abstractions of the initial links. Navigating such interrelated lattice family in order to find concepts of interest is not a trivial task due to the potentially large size of the lattices and the need to move the expert's focus from one lattice to another. In this paper, we investigate the navigation of a concept lattice family based on a query expressed by an expert. The query is defined in the terms of RCA. Thus it is either included in the contexts (modifying the lattices when feasible), or directly classified in the concept lattices. Then a navigation schema can be followed to discover solutions. Different navigation possibilities are discussed