Relationsip semi-automation in thesauri: a proposal for semantic relationship refinement based on thesaurus AGROVOC

• Autor(a) principal: Berti Junior, Decio Wey
• Data de Publicação: 2017
• Outros Autores: Soergel, Dagobert, Lima, Gercina Ângela de, Maculan, Benildes Coura Moreira dos Santos
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Informação & Informação
• Texto Completo: https://ojs.uel.br/revistas/uel/index.php/informacao/article/view/28724
• topic: Knowledge Organization System; Thesaurus; Sematic Relations; AGROVOC; Sistema de Organización del Conocimiento; Tesauro; Relaciones Semánticas; AGROVOC; Sistema de Organização do Conhecimento; Tesauros; Relações Semânticas; AGROVOC

Introduction: Thesauri are tools that contribute to information retrieval in information services like digital data bases and digital libraries. Objective: This study aims to present a quantitative analysis on the refined sematic structure of the AGROVOC Thesaurus and propose a semi-automated model for semantic relationship refinement of thesauri. Methodology: To carry out this study, we used data from the AGROVOC Thesaurus represented in a refined SKOS-XL model. This data has been qualitatively aggregated in the classification of concepts by entity types as well as the hierarchical classification of relationships done by Soergel. From this base, we conducted a quantitative analysis of the relationship types. Results: Results from the quantitative data analysis showed that the refinement of the AGROVOC Thesaurus is not yet complete. Most Related Term relationships seem to have been refined, but hierarchical relationships (Broader / Narrower) have not. Conclusion: In sum, this study demonstrates that quantitative analysis can shed light on the structure of a thesaurus and indicate areas where improvements are possible.