Corpus of Academic Czech

The Academic Czech Corpus is a complement to Phrase Bank of Academic Czech and includes only Czech-language untranslated texts published after 2010 in scientific journals indexed in Web of Science, Scopus or EBSCO databases. Another criterion is the genre of the text: only studies and review articles are included in the corpus, not, for example, reviews or conference reports. In most cases, the texts are in the final editing stage, i.e. they have not undergone final editing or proofreading. The corpus contains articles from a total of 21 titles of Czech-language scientific journals and all six disciplines of the broader Frascat Manual are represented. A more precise composition of the corpus is given in the table below. The predominance of social sciences and humanities is due to the fact that relatively few Czech-language scientific articles are published in other disciplines.

The total extent of the corpus is more than 15 million words (almost 20 million tokens) in 3,394 scientific articles. The technical processing of the corpus is based on the corpora of the SYN series. The main difference with the SYN series is that the documents here correspond to individual articles, not numbers. In addition, documents (articles) are further divided into individual sections (<div>) corresponding to text sections with an explicit class designation, which takes on the values introduction, discussion, conclusion/ / and unknown. This breakdown was obtained by heuristic procedures and is therefore not always reliable. Metadata (authors, article title, number, year of publication, etc.) is available for all documents, which has undergone extensive manual revision. The lemmatization and morphological tagging of the corpus correspond to SYN2020. The author's team would like to thank the editors of the journals included in the corpus, without whose support the Corpus of Academic Czech could not have been created. ====== How to cite Corpus of Academic Czech ======