Akademska digitalna zbirka SLovenije - logo
VSE knjižnice (vzajemna bibliografsko-kataložna baza podatkov COBIB.SI)
  • sloWnet : construction and corpus annotation
    Fišer, Darja, 1978- ; Erjavec, Tomaž, 1960-
    This paper presents a wordnet for Slovene which was created semi-automaticallywith a combination of approaches and multilingual resources,in particular a bilingual dictionary, a parallel corpus and ... Wikipedia. Analysis of the results shows that the dictionary approach yields agood core wordnet but requires substantial manual editing due to a lack of automatic word-sense disambiguation. This was successfully improved with the corpus approach which, however, was limited to single-word literals. The last approach, based on Wikipedia, was only used for domain-specific monosemous terms, and can deal with multi-word literals and therefore usefully complements the previous two approaches. The created sloWNet was then used to semantically annotate a corpus for Slovene: one hundred high frequency nouns were annotated in a corpus of 100,000 words. The paper reports on the method and results of this manual annotation. Both the Slovene wordnet and annotated corpus are to be publicly available.
    Vir: Principles, construction and application of multilingual wordnets : proceeding [!] of the 5th Global Wordnet Conference (Str. 177-183)
    Vrsta gradiva - prispevek na konferenci
    Leto - 2010
    Jezik - angleški
    COBISS.SI-ID - 42066018

Vnos na polico