Experiencia de clasificación automática de documentos sobre Ciencias de la Vida y Biomedicina obtenidos del Web of Science

Abstract

This article provides an experience in the development and proof of a classification algorithm that automatically assigns a theme to a document. The Naïve Bayes Multinomial classification was used to automatically analyze the correlation between the themes of research in Life Sciences and Biomedicine, and the result of a corpus of 10 167 articles recuperated from the Web of Science (WoS). A proof of the performance of the algorithm was applied to 5 581 reviews for measuring its precision, exhaustivity and accuracy. The results show its usefulness in the organization of information respect to the automatic assignation of themes to the documents obtained in a digital repository or a bibliographic data base. The algorithm proposed can be utilized as an alternative to the traditional methods of classification of documents in a specific area of knowledge; this will allow the creation of specialized services oriented to the development of computational services that support the digital and electronic information management.