Extraction probabiliste de formes SHACL à l'aide d'algorithmes évolutionnaires
Abstract
SHACL is the W3C recommendation to represent constraints on RDF graphs called shapes, used to validate RDF data. Creating shapes capturing domain constraints is a very tedious task. That is why recent work addresses the problem of automatic shapes generation. In this context, we propose an approach combining grammatical evolution and a probabilistic framework to generate candidate SHACL shapes and validate RDF graphs against candidate shapes. The approach is validated by mining SHACL shapes representing association rules from an RDF dataset. The results show the potential and generality of the proposed approach.