The Reference Corpus of the Contemporary Romanian Language (CoRoLa)

The Reference Corpus of the Contemporary Romanian Language (CoRoLa) Verginica Barbu Mititelu, Dan Tufiș, Elena Irimia Romanian Academy Research Institute for Artificial Intelligence 13 13 Septembrie Road, Bucharest 050711, Romania {vergi, tufis, elena}@racai.ro Abstract We present here the largest publicly available corpus of Romanian. Its written component contains 1,257,752,812 tokens, distributed, in an unbalanced way, in several language styles (legal, administrative, scientific, journalistic, imaginative, memoirs, blogposts), in four domains (arts and culture, nature, society, science) and in 71 subdomains. The oral component consists of almost 152 hours of recordings, with associated transcribed texts. All files have CMDI metadata associated. The written texts are automatically sentence- split, tokenized, part-of-speech tagged, lemmatized; a part of them are also syntactically annotated. The oral files are aligned with their corresponding transcriptions at word-phoneme level. The transcriptions are also automatically part-of-speech tagged, lemmatised and syllabified. CoRoLa contains original, IPR-cleared texts and is representative for the contemporary phase of the language, covering mostly the last 20 years. Its written component can be queried using the KorAP corpus management platform, whereas the oral component can be queried via its written counterpart, followed by the possibility of listening to the results of the query, using an in- house tool. Keywords: Romanian, reference corpus, annotation 1. Introduction Language resources in the form of large corpora have been being created for more and more languages. We present here the results of a four-year project focused on the creation of a big corpus for contemporary Romanian language, called CoRoLa (corola.racai.ro). This has not been a singular effort: smaller previous or parallel projects (ANVSIB – http://speed.pub.ro/anvsib, SSPR – http://dev.racai.ro/ti/wordpress/, PARSEME – https://typo.uni-konstanz.de/parseme/) contributed to this project outcomes. In its turn, CoRoLa will contribute to other projects, larger ones, going beyond the national level (DruKoLA – http://www1.ids-mannheim.de/ direktion/kl/projekte/drukola.html) and turning European. The CoRoLa corpus is now the reference corpus for contemporary Romanian. It is the largest one, containing 1,257,752,712 tokens for the written component and almost 152 hours of recordings for the oral component (the detailed structure is presented below). The texts cover all language styles, four major domains for which 71 subdomains were defined, thus ensuring a wide vocabulary coverage. The corpus can be reliably used as a basis for the creation of dictionary entries, grammar studies, other language reference materials, as well as for training and testing algorithms and systems for language processing. CoRoLa has been jointly developed, as a priority project of the Romanian Academy, by two institutions: “Mihai Drăgănescu” Research Institute for Artificial Intelligence (from Bucharest) and the Institute of Computer Science (from Iași). Being located in different geographical and cultural regions of the country, the two partners could more easily contact texts providers from their vicinity, as unmediated, face-to-face contact and negotiations proved necessary for agreeing upon a protocol of collaboration after the correct understanding by our texts providers of the way texts are to be processed and further exploited when part of the corpus. 2. Related work National corpora have been created for many languages: American English …

Content truncated to 500 words