Propuesta y evaluación de un método extractivo de generación de resumenes en el ámbito biomédico basado en conceptos

Abstract

The methods for automatic summarization generation based in extractive techniques have widely shown its utility for his adaptability and efficiency in the manner of response time at any kind of application domain. In Biomedical field are numerous the research results about the overload information and the need of application of efficient recovery and summarization methods for the proper use of evidence based medicine. In this context we are going to present a proposal of methodology for automatic summarization based on structured knowledge and graph's use, applying similarity methods between phrases and considering concepts appearance frequency. Finally, a methodology’s evaluation is made to compare with other methods using the ROUGE tool and analyzing their results. Although the size of the evaluation set doesn’t allow extracting noteworthy conclusions, the results collected are enough promising to trust in the effectiveness of the proposal handed in.