Homonymie mezi apelativy a proprii jako problém automatické morfologické analýzy češtiny

Title in English [Homonymy among Czech common and proper nouns as the problem of automatic morphological analysis


Type Article in Periodical
Magazine / Source Acta onomastica
MU Faculty or unit

Faculty of Arts

Web Acta Onomastica
Keywords tokenisation; lemmatisation; disambiguation; corpus linguistics
Description The aim of this paper is a corpus-based analysis of one type of Czech proper nouns (type Zubří). We will argue that adequate annotation (lemmatisation and morphological tagging) of proper nouns type Zubří depends on several circumstances: 1) coverage of the dictionary of the automatic analyzer; 2) accurate description of the variability of inflexion forms; 3) not trivial disambiguation of numerous homonym word-forms. We believe that while satisfying the first two conditions is possible, adequate disambiguation goes beyond the possibilities of automatic morphological analysis.
Related projects: