We tackled here the question of whether RCA brings some effective scope extension to the realm of FCA, given that FCA is at its core.
Relational Concept Analysis (RCA) was designed as an extension of Formal Concept Analysis (FCA) to multi-relational datasets, such as the ones drawn from Linked Open Data (LOD) by the type-wise grouping of the resource into data tables.
RCA has been successfully applied to practical problems of AI such as knowledge elicitation, knowledge discovery from data and knowledge structuring.
A crucial question, yet to be answered in a rigorous manner, is to what extent RCA is a true extension of FCA, i.e. reveals concepts that are beyond the reach of core FCA even using a suitable encoding of the original data.
We show in this article that the extension is effective: RCA retrieves all concepts found by FCA as well as many further ones.
Michael Wajnberg, Petko Valtchev (Université du Québec à Montréal, Canada) and Mario Lezoche, Alexandre Blondin-Massé, Hervé Panneto (Université de Lorraine, France)