A Tagged Corpus for Automatic Labeling of Disabilities in Medical Scientific Papers

A Tagged Corpus for Automatic Labeling of Disabilities in Medical Scientific Papers Carlos Valmaseda, Juan Martinez-Romo, Lourdes Araujo NLP&IR Group, Universidad Nacional de Educaci´on a Distancia (UNED) 28040 Madrid, Spain carlosvalmaseda@gmail.com, juaner@lsi.uned.es, lurdes@lsi.uned.es Abstract This paper presents the creation of a corpus of labeled disabilities in scientific papers. The identification of medical concepts in documents and, especially, the identification of disabilities, is a complex task mainly due to the variety of expressions that can make reference to the same problem. Currently there is not a set of documents manually annotated with disabilities with which to evaluate an automatic detection system of such concepts. This is the reason why this corpus arises, aiming to facilitate the evaluation of systems that implement an automatic annotation tool for extracting biomedical concepts such as disabilities. The result is a set of scientific papers manually annotated. For the selection of these scientific papers has been conducted a search using a list of rare diseases, since they generally have associated several disabilities of different kinds. Keywords: corpus, annotation, medical concepts, disabilities 1. Introduction The study of the relationships between different elements of the biomedical domain is essential for further progress in the area. Great efforts are being devoted to identify some of these relationships, such as interactions between proteins, genes-diseases associations or adverse drug effects. The way to deal with these problems usually involves the identi- fication of some of these relationships by experts. For both addressing this problem, and finding specialized terminol- ogy related to a specific aspect of the domain it is essential the annotation of corresponding concepts such as diseases, genes, proteins, etc. In this paper we address the annotation of a type of concept that is not collected in previous works, at least not specif- ically. This is the identification of expressions related to disabilities. While some disabilities are included among the symptoms of some biomedical domain ontologies, they are only a few cases and their identification needs to be tackled in a specific way. In this paper this problem is addressed, that although shares some aspects with the annotation of concepts in the biomedical domain, also presents particular aspects, since references to disabilities can be more freely expressed that references to diseases, genes, proteins, etc. References to disabilities supports all kinds of syntactic, morphological, and semantic variations. For instance, for the same disabilities, the following variants could be found: • I can not move my left leg • Mobility limitations in lower limbs • The leg does not respond to the patient Therefore in this context, to apply natural language pro- cessing (NLP) techniques becomes more relevant. Taking into account that there are currently no resources with which to evaluate a possible system that tries to de- tect these disabilities, we proposed to develop a corpus of documents which includes several disabilities. As experimental framework, we focused on the disabili- ties associated with rare diseases (RD) for several reasons. On the one hand it is a problem of great importance given …

Content truncated to 500 words