Impact précoce du poids des balises pour la recherche d'information ciblée

Mathias Géry (1) , Christine Largeron (1) , Franck Thollard (1)

1 LHC - Laboratoire Hubert Curien

This paper addresses the integration of XML tags in terms weighting function for focused XML Information Retrieval (IR). Our model allows to consider a certain kind of structural information: tags that represent logical structure (title, section, paragraph, etc.) as well as tags related to formatting (bold, italic, center, etc.). We take into account the tags influence by estimating the probability that tags distinguishe relevant terms. Then, these weights are integrated in terms weighting function. Experiments on a large collection during INEX 2008 XML IR evaluation campaign showed improvements on focused retrieval.

Mots clés: Modèle probabiliste de document; Recherche d'information structurée; XML; Balises; Pondération

Domaines: Recherche d'information [cs.IR]

Mathias Géry, Christine Largeron, Franck Thollard. Impact précoce du poids des balises pour la recherche d'information ciblée. Conférence en Recherche d'Information et Applications, May 2009, Presqu'île de Giens, France. pp 333-348. ⟨ujm-00385200⟩