Morphosyntactic structure of terms in Basque for automatic terminology extraction
Rubén Urizar, Nerea Ezeiza, Iñaki Alegria · 2000
This paper describes the morphosyntactic patterns of technical terms in Basque and presents an architecture for a term-extracting tool. As Basque is a highly inflected agglutinative language, part-of-speech information is not enough to define term patterns. The use of morphological and syntactic information is essential to reduce considerably the number of structures. For example, a noun, an adverbial, a postpositive adjectival, the nominal form of a verb and even a determiner in the genitive case may work as a prepositive adjective; however, they all share the same syntactic function. Therefore, for the term-extracting tool to perform properly, the texts must be morphosyntactically analysed and disambiguated. Then a shallow syntactic parser will identify the previously described patterns.