Skip to Main content Skip to Navigation
Journal articles

REISA : Enrichissement contrôlé de bases de connaissances à partir de documents annotés

Abstract : Many annotation tools have been proposed for the automatic annotation of documents but the annotation of semantic relations between concept instances remains a challenging task when structures and vocabularies are heterogeneous. We propose an approach, called REISA, which allows to enrich knowledge bases with property instances using semistructured documents annotated with concept instances. The approach relies on a neighborhood distance between annotated domain entities in the document and perform a semantic control of the extracted facts using axioms and facts of the knowledge bases. The experiments show that REISA allows to extract property instances with interesting precision values reaching 76.5%.
Document type :
Journal articles
Complete list of metadatas

https://hal-supelec.archives-ouvertes.fr/hal-01061253
Contributor : Elodie Dubrac <>
Submitted on : Friday, September 5, 2014 - 2:04:38 PM
Last modification on : Wednesday, September 16, 2020 - 4:57:34 PM

Identifiers

Collections

Citation

Nacéra Bennacer Seghouani, Nathalie Pernelle, Yassine Mrabet. REISA : Enrichissement contrôlé de bases de connaissances à partir de documents annotés. Revue des Sciences et Technologies de l'Information - Série RIA : Revue d'Intelligence Artificielle, Lavoisier, 2014, 28 (2-3), ⟨10.3166/ria.28.297-320⟩. ⟨hal-01061253⟩

Share

Metrics

Record views

361