The slWaC Corpus of the SloveneWeb

Abstract

The availability of large collections of text (language corpora) is crucial for empirically supported linguistic investigations of various languages; however, such corpora are complicated and expensive to collect. In recent years corpora made from texts on the World Wide Web have become an attractive alternative to traditional corpora, as they can be made automatically, contain varied text types of contemporary language, and are quite large. The paper describes version 2 of slWaC, a web corpus of Slovene containing 1.2 billion tokens. The corpus extends the first version of slWaC with new materials and updates the corpus compilation pipeline. The paper describes the process of corpus compilation with a focus on near-duplicate removal, presents the linguistic annotation, format and accessibility of the corpus via web concordancers. It then investigates the content of the corpus using the method of frequency profiling, by comparing its lemma and part-of-speech annotations with three corpora: the first version of slWaC, with Gigafida, the one billion word reference corpus of Slovene, and KRES, the hundred million word reference balanced corpus of Slovene.