Évaluation des améliorations de prédiction d'hospitalisation par l'ajout de connaissances métier aux dossiers médicaux
In EGC 2019, vol. RNTI-E-35, pp.249-254
The knowledge available through electronic medical records (EMR) themselves remains limited by the fact the features used by a machine learning algorithms from a text alone do not contain all the implicit information known by a domain expert. We propose and evaluate the ontological augmentations of features extracted from textual information from EMRs on several machine learning algorithms to predict hospitalization.