dc.contributor.advisor | Nytrø, Øystein | nb_NO |
dc.contributor.advisor | Slaughter, Laura | nb_NO |
dc.contributor.advisor | Helgeland, Jon | nb_NO |
dc.contributor.author | Schmidt, Jarl Erik | nb_NO |
dc.date.accessioned | 2014-12-19T13:38:09Z | |
dc.date.available | 2014-12-19T13:38:09Z | |
dc.date.created | 2012-02-02 | nb_NO |
dc.date.issued | 2011 | nb_NO |
dc.identifier | 489214 | nb_NO |
dc.identifier | ntnudaim:5770 | nb_NO |
dc.identifier.uri | http://hdl.handle.net/11250/252756 | |
dc.description.abstract | This thesis describes an experiment to create an Information Extraction tool capable of identifying disease- and intervention-related concepts in Norwegian-language patient narratives. Due to the unavailability of training data, the challenge was to find a novel approach using only clinical knowledge resources.The approach was to create an ontological representation of the Norwegian version of the International Classification of Disease (10th revision), and design an algorithm that uses concept matching and structural knowledge of the classification to choose the most likely codes.Evaluation on a minimal annotated text corpora shows that while the recall is low, the approach shows some degree of promise. | nb_NO |
dc.language | eng | nb_NO |
dc.publisher | Institutt for datateknikk og informasjonsvitenskap | nb_NO |
dc.subject | ntnudaim:5770 | no_NO |
dc.subject | MIT informatikk | no_NO |
dc.subject | Systemarbeid og menneske-maskin-interaksjon | no_NO |
dc.title | Ontology-based extraction of clinical concepts from narrative text | nb_NO |
dc.type | Master thesis | nb_NO |
dc.source.pagenumber | 93 | nb_NO |
dc.contributor.department | Norges teknisk-naturvitenskapelige universitet, Fakultet for informasjonsteknologi, matematikk og elektroteknikk, Institutt for datateknikk og informasjonsvitenskap | nb_NO |