Recherche documentaire par titrage automatique

In this paper, we propose a system in order to facilitate the information retrieval in a set of textual documents. Our approach is based on the automatic titling (and subtitling). This last one is crucial, for example, for the issue of web pages accessibility (W3C standard). Our process of automatic titling consists in extracting relevant noun phrases from texts. These ones can represent a title/subtitle of the document. An original approach combining statistical criteria and placement of the noun phrases in the text allows to propose titles and relevant subtitles. So, the user can have an outline of all the subjects evoked in a mass of documents, and easily find the information he was looking for. An evaluation on real data shows that the solutions given by our automatic titling approach are relevant. MOTS-CLÉS : Syntagmes nominaux, titrage automatique, statistiques.