Euskarazko ezeztapenaren tratamendu automatikorakoazterketa

  1. Begoña Altuna Diaz
  2. Maxux Aranzabe Urruzola
  3. Arantza Diaz de Ilarraza Sanchez
Livre:
II. Ikergazte nazioarteko ikerketa euskaraz: Kongresuko artikulu-bilduma.
  1. Iñaki Alegria (ed. lit.)
  2. Ainhoa Latatu (ed. lit.)
  3. Miren Josu Omaetxebarria (ed. lit.)
  4. Patxi Salaberri (ed. lit.)

Éditorial: Udako Euskal Unibertsitatea, UEU = Universidad Vasca de Verano

ISBN: 978-84-8438-627-8 978-84-8438-628-5

Année de publication: 2017

Titre du volume: Giza Zientziak eta Artea

Tomo: 1

Volumen: 1

Pages: 127-134

Congreso: Ikergazte. Nazioarteko Ikerketa Euskaraz (2. 2017. Iruñea)

Type: Communication dans un congrès

DOI: 10.26876/IKERGAZTE.II.01.17 DIALNET GOOGLE SCHOLAR lock_openAccès ouvert editor

Résumé

Natural language processing aims to the information extraction and analysis. In this paper, we presentthe analysis and annotation of negation in Basque, as a complement to the temporal informationprocessing, since negation affects event factuality resolution and event role definition. We have analysedthe background and we have studied how negation is expressed in Basque and we have normalisedthat information. In order to evaluate the normalisation scheme and to create a corpus for machine learning, we have conducted a manual annotation experiment. Based on the experiment results,we have analysed the errors to measure the difficulty of the task and to improve the annotation guidelines.