Gesellschaft für Informatik e.V.

Lecture Notes in Informatics

Natural Language Processing and Information Systems, 8th International Conference on Applications of Natural Language to Information Systems, June 2003, Burg (Spreewald), Germany. P-29, 200-213 (2003).

GI, Gesellschaft für Informatik, Bonn


Antje Düsterhöft (ed.), Bernhard Thalheim (ed.)

Copyright © GI, Gesellschaft für Informatik, Bonn


Linguistic resource for NLP: ask for "die drei musketiere" and meet "LES trois mousquetaires"

Odile Piton , Thierry Grass and Denis Maurel


Our work concerns proper names or "named entities" in NLP, in a multilingual context and in the spirit of the action "Technolangue" launched in France by the Ministry for Research, the Ministry for the Culture and the Communication and the Ministry for the Economy, Finances and Industry in 2002, with the purpose of production, validation and diffusion of linguistic resources. Our objective is to create a multilingual electronic dictionary intended to record such terms in several languages as well as the bonds between them, and We know that toponyms follow strict rules imposed by international or national organizations (UNGEGN) intended to standardize them. This is not true for the other categories of proper names. In order to create an operational tool that is able to help translation, we must study the proper names, their grammar and the semantic relations between them. We propose ontology based on a taxonomy inspired by the works of Bauer and others. We took as a starting point the work of Mel'cuk on the lexical functions as well as work of G. Miller on the WorldNet system. The basic model includes semantic or lexical relations. These relations make it possible to locate a proper name in a lexical network and to provide responses on request.

Full Text: PDF

GI, Gesellschaft für Informatik, Bonn
ISBN 3-88579-358-X

Last changed 04.10.2013 17:57:46