A bayesian classifier for the recognition of the impersonal occurrences of the 'it' pronoun
Résumé
This paper presents a new system that makes the distinction between the impersonal and anaphoric occurrences of the 'it' pronoun. Compared with the state of the art methods, our system relies on the same types of linguistic knowledge but performs better. We argue that this is due to the bayesian model on which it is based: it enables to combine various pieces of knowledge and to exploit even unreliable ones in the process of pronoun occurrence classification.
Domaines
Traitement du texte et du document
Origine : Fichiers produits par l'(les) auteur(s)