Skip to Main content Skip to Navigation
Master Thesis

Extraction d'expressions polylexicales sur corpus arboré

Abstract : This document describes a method for extracting multiword expressions from syntactically analyzed corpora. These expressions are still a main issue for NLP applications. Their automatic extraction, and appropriate description, remain largely unsolved problems. This approach most notably allows the extraction of expressions composed of more than two words, and focuses on the issue of nested recurrent structures. It also yields a finegrained morphosyntactic description of the extracted units, including syntactic relations, possible word orders, contiguity, inflection or determination, which we think is necessary to a proper encoding of these expressions in lexicons, for applications such as information extraction or machine translation.
Complete list of metadatas

Cited literature [31 references]  Display  Hide  Download

https://dumas.ccsd.cnrs.fr/dumas-00704873
Contributor : Uga - Bulles <>
Submitted on : Wednesday, June 6, 2012 - 2:12:51 PM
Last modification on : Wednesday, July 15, 2020 - 1:06:02 PM

Identifiers

  • HAL Id : dumas-00704873, version 1

Citation

Julien Corman. Extraction d'expressions polylexicales sur corpus arboré. Sciences de l'Homme et Société. 2012. ⟨dumas-00704873⟩

Share

Metrics

Record views

468

Files downloads

1556