Title :
A Lightweight Pronoun Resolution Algorithm for French Corporate Texts
Author :
Tardif, Olivier ; Ratte, Sylvie
Author_Institution :
Univ. de Provence, Aix-en-Provence
Abstract :
In this paper we present COREFF, a pronoun resolution algorithm implemented as a GATE plugin, composed of a small set of rules that follow from observations made on a corpus of French corporate texts, and requiring very little linguistic processing besides what is already available for GATE. We show that its performances are rather good for such a lightweight system, but that it shows, however, a strong dependency to the corpus it was developed for. We compare its results with the ones obtained by a similar algorithm for English on two different corpora, and give a brief description of the errors.
Keywords :
computational linguistics; natural language processing; text analysis; COREFF lightweight pronoun resolution algorithm; French corporate text; GATE plugin; linguistic processing; natural language processing algorithm; Data mining; Government; Knowledge management; Machine learning; Machine learning algorithms; Natural languages; Performance evaluation; Project management; Technology management; Unified modeling language;
Conference_Titel :
Advanced Computer Theory and Engineering, 2008. ICACTE '08. International Conference on
Conference_Location :
Phuket
Print_ISBN :
978-0-7695-3489-3
DOI :
10.1109/ICACTE.2008.215
