Generation of Feature Vectors for Identifying Medical Entities in Spanish

Abstract

Natural Language Processing (NLP) encompasses a range of high-impact techniques to enable computers to interact with humans more naturally. One such technique is the extraction of entities, which allows computers to identify relevant information within a text. This paper presents a methodology for recognizing medical entities within texts written in Spanish. The methodology combines syntactic, semantic and contextual features at the word level. The main aim of the feature-based approach is to identify drug, anatomy, and disease entities. A training evaluation was conducted on two machine learning algorithms, with an precision of 98% on an external set. In addition, an precision check was performed for each medical class.