Approches endogène et exogène pour améliorer la segmentation thématique de documents

Topic segmentation was addressed by a large amount of work from which it is not easy to draw conclusions, especially about the need for knowledge. In this article, we propose in the same framework two methods for improving the results of a topic segmenter based on lexical reiteration. The first one is endogenous and exploits the distributional similarity of the words of a document for discovering its topics. These topics are then used to facilitate the detection of topical similarity between discourse units. The second approach achieves the same goal by relying on an external resource, that is a network of lexical co-occurrences built from a large corpus. These two approaches are also combined. An evaluation of these approaches and their combination is performed in a reference framework and shows the interest of this combination. MOTS-CLÉS : analyse thématique, segmentation thématique, découverte de thèmes, cooccurrences lexicales.