Using Formal Concept Analysis for Discovering Knowledge Patterns

Amine Mohamed Rouane Hacene (1) , Marianne Huchard (2) , Amedeo Napoli (3) , Petko Valtchev (1)

1 Département d'informatique [Montréal]
2 MAREL - Models And Reuse Engineering, Languages
3 ORPAILLEUR - Knowledge representation, reasonning

Design patterns are used in software engineering for guiding code design: they play the role of models to be followed for producing code of better quality. In the same way, knowledge patterns are intro- duced in knowledge engineering as ontology components that can be used as models and reused as ontology design patterns (ODPs) in ontology engineering. The representation of domain knowledge is not an easy task and this is actually the case for knowledge patterns. Accordingly, we present in this paper the use of both Formal Concept Analysis (FCA) and Relational Concept Analysis (RCA) for designing compact concept lattices that can be re-engineered as ODPs. Starting with a simple ex- ample, it is shown how to derive conceptual and relational abstractions that can be considered as knowledge patterns, following and adapting the meta-modeling activity in software engineering. This paper aims at showing that FCA and RCA are e

Mots clés: Formal Concept Analysis; Relational Concept Analysis; Ontology; Design pattern; Ontology pattern

Domaines: Intelligence artificielle [cs.AI]

Amine Mohamed Rouane Hacene, Marianne Huchard, Amedeo Napoli, Petko Valtchev. Using Formal Concept Analysis for Discovering Knowledge Patterns. CLA: Concept Lattices and their Applications, Oct 2010, Sevilla, Spain. pp.223-234. ⟨lirmm-00531802⟩