The hundred million word corpus Kres was compiled in order to support the Gigafida corpus with a more balanced source of contemporary Slovene. The paper present the taxonomic and text-type composition of both corpora and the topic model of the Kres corpus. In order to illustrate the differences and similarities between Gigafida and Kres, which are appropriate corpus sources for a new dictionary of Slovene, we selected a set of salient nouns and analysed their typical contexts in terms of Word sketches as obtained from both corpora.
|
