%0 Journal Article
%T Spotting scientific and technical specialization in biomedical documents using morphological clues D¨¦tection de la sp¨¦cialisation scientifique et technique des documents biom¨¦dicaux gra ce aux informations morphologiques
%A Jolanta Chmielik
%A Natalia Grabar
%J Traitement Automatique des Langues
%D 2012
%I Association pour le Traitement Automatique des Langues (ATALA)
%X Distinction of the specialization level of the health documents on Internet is an important indication, especially when documents are read by non expert users such as patients. Indeed, a high technicity of documents impedes the patients to understand the content and may have a negative consequence on their health care process and on their communication with medical doctors. When medical portals propose such a distinction, it is obtained further to a human categorisation. We propose an automatic categorization of health documents according to their specialization. We exploit morphological information obtained thanks to the morphological analysis of lexems. The evaluation shows that precision, recall and f-measure are often higher than 90%.
%K medical documents
%K specialization
%K supervised machine learning
%K contructional morphology
%K semantics
%U http://www.atala.org/IMG/pdf/5-Chmielik-TAL52-2-2011.pdf