La visualisation de données relationnelles au service de la recherche d'informations

Within the framework of the information retrieval, the restitution of the documents is carried out according to their calculated relevance score. They correspond to a precise request. However questions arise as for the representation of documents and requests like their mapping. The graph is used as data visualization, without requiring mathematical knowledge. This article presents the contributions for the information visualization. We stick to the design process of a visualization and analysis tool developed in our laboratory IRIT by the SIG/EVI team: VisuGraph prototype. In this context, we propose new approaches for the information networks visualization, in order to extract easily the most relevant documents. We develop this approach and we illustrate our contribution by real examples. MOTS-CLES : Recherche d’information, pertinence, dessin de graphe, visualisation des résultats de recherches, analyse relationnelle, réseaux sémantiques.