Search

This paper is about the creation of an electronic dictionary of sports terminology, which will be used by computers in natural language processing systems. This dictionary is based on the Greek version of a multilingual sports terminology database, which was processed within the framework of the Euradic project as part of the Technolangues project. The terms were first translated into Greek. The next step was to transfer the database entries to the electronic terminology dictionary including verbs, simple and compound nouns, adjectives, simple and compound adverbs, initialisms, frozen expressions. For the creation of this dictionary, the computational linguistics team of the Natural Language Processing Unit, Aristotle University of Thessaloniki, and Gaspard Monge Institute, University of Marne-la-Vallée used the same codification and automatic inflection programmes that were used for the creation of their general electronic dictionaries. The research was conducted by the said team, whose studies aim at a detailed and formalised description of Modern Greek, the final objective being the recognition of linguistic data by natural language processing systems.

(En grec) electronic version (13 pp.); International audience; During the process of enrichment of an au tomatic processing system in Mod ern Greek, we confronted the problem of ambigu ity between ad verbs and ad jectives with nominal use. The object of this stu d y is the retrieval of these ambiguities on the one hand , and the efforts to eliminate them on the other hand . We were based on the theoretical principles of Z. S. Harris as well as on the stu d ies of Mau rice Gross, who proposed the complete and systematic d escription of each language, aiming at the au tomatic text processing. In particu lar, the basis was a corpu s of 6.000 simple ad verbs and a corpu s of ~2.000 ad jectives with nominal u se. The verification of the analysis performed , was realised by UNITEX, an au tomatic processing system that was d eveloped in LADL (Laboratoire d 'Au tomatique et Documentaire Linguistiqu e) of the University of Marne-la-Vallee. During the research, we have come to the following conclusions: - There is ambiguity when an adjective with nominal use in neuter gender is found in the nominative, accusative and vocative of plural. For example: 1.0. / (daily/specially), adverb 1.1. / (everyday clothes/private lessons), adjective with nominal use (She is wearing her everyday clothes) (He offers private lessons) (He comes on a daily basis and meets me) (He is specially interested in literature) In order to resolve such ambiguities, there are the following solutions: - finite state automata, - creation of syntactical-semantic analysers, - classes of objects.