Building Croatian WordNet

Abstract

This paper reports on the prototype Croatian WordNet (CroWN). The resource has been collected by translating BCS1 and 2 from English, but also by usage of machine readable dictionary of Croatian language which was used for automatical extraction of semantic relations and their inclusion into CroWN. The paper presents the results obtained, discusses some problems encountered along the way and points out some possibilities of automated acquisition and populating synsets and their refinement in the future. In the second part the paper discusses the lexical particularities of Croatian, which are also shared between other Slavic languages (verbal aspect and derivation patterns), and points out the possible problems during the process of their inclusion in CroWN.