Une évaluation de l'impact des types de textes sur la tâche de segmentation thématique

Abstract : This paper aims to contribute to a better definition of the requirements of the text segmentation task, by stressing the need for taking into account the types of texts that can be appropriately considered. Our hypothesis is that while TS is indeed relevant to analyse texts with a thematic organisation, this task is ill-fitted to deal with other modes of text organisation (temporal, rhetorical, etc.). By comparing the performance of a TS system on two corpora, with either a "strong" or a "weak" thematic organisation, we show that TS is sensitive to text types.
Document type :
Conference papers
Complete list of metadatas

https://hal.inria.fr/inria-00511605
Contributor : Philippe Muller <>
Submitted on : Wednesday, August 25, 2010 - 3:01:09 PM
Last modification on : Friday, June 14, 2019 - 6:31:00 PM

Identifiers

  • HAL Id : inria-00511605, version 1

Citation

Clémentine Adam, Philippe Muller, Cécile Fabre. Une évaluation de l'impact des types de textes sur la tâche de segmentation thématique. Traitement Automatique des Langues Naturelles - TALN 2010, Jul 2010, Montréal, Canada. ⟨inria-00511605⟩

Share

Metrics

Record views

477