BioRo: The Biomedical Corpus for the Romanian Language

BioRo: The Biomedical Corpus for the Romanian Language Maria Mitrofan and Dan Tufis¸ Research Institute for AI ”Mihai Dr˘ag˘anescu”, Romanian Academy ”Calea 13 Septembrie”, Bucharest 050711, Romania {maria, tufis}@racai.ro Abstract The biomedical domain provides a large amount of linguistic resources usable for biomedical text mining. While most of the resources used in biomedical Natural Language Processing are available for English, for other languages including Romanian the access to language resources is not straight-forward. In this paper, we present the biomedical corpus of the Romanian language, which is a valuable linguistic asset for biomedical text mining. This corpus was collected in the contexts of CoRoLa project, the reference corpus for the contemporary Romanian language. We also provide informative statistics about the corpus, a description of the data-composition. The annotation process of the corpus is also presented. Furthermore, we present the fraction of the corpus which will be made publicly available to the community without copyright restrictions. Keywords: BioNLP; Romanian; Corpora; Specialized Domain; 1. Introduction Natural language processing (NLP) technologies have been used to extract useful information from different types of biomedical texts, such as: biomedical literature, research papers, medical school lecture notes, clinical notes, dis- charge summaries, clinical practice guidelines, etc. In order to be able to obtain relevant knowledge form tex- tual data, high-quality resources (i.e. corpora, lexica, ter- minologies, tesauri etc.) are needed. At international level various biomedical textual resources have been developed, but most of them are available in English (e.g.GENIA cor- pus (Kim et al., 2003), AnEM corpus (Ohta et al., 2012), CellFinder corpus (Neves et al., 2012), etc.). In many countries where the official language is not En- glish, there is a technical barrier in using Natural Language Processing in the biomedical domain (bioNLP) due to the fact that textual resources are more scarce. Nevertheless, a significant progress has been made over the past decades thanks to the contribution of various active NLP communi- ties. For example, for French the language Zweigenbaum et al. (2005) created the ”Unified Medical Lexicon for French” (UMLF), a reference resource for bioNLP. Another im- portant linguistic resource is Corpus Medical du Centre de Recherche en Terminologie et Traduction (CMCRTT), which is a monolingual French corpus for bioNLP and it is publicly available both as plain or POS tagged text (Neveol et al., 2014). For the Bulgarian language important efforts have been made in developing resources usable for various NLP tasks. Boytcheva et al. (2009) collected a biomedical corpus con- taining 6400 words, 2000 of them are part of the Bulgarian medical terminology. Swedish is another language which has important resources for bioNLP. In 2012, Velupillai (2012) created an anno- tated gold standard of medical records, used for terminol- ogy management and linguistic explorations. Also, a nega- tion and clinical uncertainty taxonomy schema was pro- posed for Swedish language and was mapped to an English annotation schema (Mowery et al., 2012). The Romanian language is an under-resourced language, regarding resources available for bioNLP. Therefore, con- siderable efforts are carried on in …

Content truncated to 500 words