A Proposal for Classifying the Content of the Web of Data Based on FCA and Pattern Structures.

This paper focuses on a framework based on Formal Concept Analysis and the Pattern Structures for classifying sets of RDF triples. Firstly, this paper proposes a method to construct a pattern structure for the classification of RDF triples w.r.t. domain knowledge. More precisely, the poset of classes representing subjects and objects and the poset of predicates in RDF triples are taken into account. A similarity measure is also proposed based on these posets. Then, the paper discusses experimental details using a subset of DBpedia. It shows how the resulting pattern concept lattice is built and how it can be interpreted for discovering significant knowledge units from the obtained classes of RDF triples.