Uso de Grafos de Conceptos para la Generación Automática de Resúmenes en Biomedicina

One of the main problems in research on automatic summarization is the inaccurate semantic interpretation of the source, which is reflected in the deficiencies shown by the resulting summaries. Using specific domain knowledge, as that supplied by ontologies, can considerably alleviate the problem. In this paper, we introduce an ontology-based extractive method for summarization. It is based on mapping the text to concepts in the ontology and representing the document as a scale-free graph. To assess the importance of the sentences we compute the centrality of their concepts in the text. We have applied our approach to summarize scientific biomedical literature, taking advantage from free resources as UMLS. Preliminary empirical results are presented and pending problems are identified.