Vis enkel innførsel

dc.contributor.advisorAamodt, Agnarnb_NO
dc.contributor.advisorFjellheim, Roarnb_NO
dc.contributor.authorOpheim, Madsnb_NO
dc.date.accessioned2014-12-19T13:41:27Z
dc.date.available2014-12-19T13:41:27Z
dc.date.created2014-09-30nb_NO
dc.date.issued2014nb_NO
dc.identifier751078nb_NO
dc.identifierntnudaim:8490nb_NO
dc.identifier.urihttp://hdl.handle.net/11250/253748
dc.description.abstractCreek is a system for diagnosis and problem solving. It makes heavily use of general knowledge, and is intended for open and weak theory domains. The system uses this general knowledge in combination with experiences from previous cases to solve new problems.We have implemented the core parts of Creek inside the Colibri studio framework, and shown that it is both possible and realistic to create a thorough Creek implementation. This has resulted in a system named Colibreek.In this implementation, we have replaced the knowledge representation in Creek, switching from the original frames to description logics. This has proven successful, and promising for further development.We have tested our system on the turbine sensor domain, and shown how our system can be used for condition monitoring on sensor-monitored equipment such as turbines. By doing this, we have also shown how the system can solve real problems people, companies and organisations have.nb_NO
dc.languageengnb_NO
dc.publisherInstitutt for datateknikk og informasjonsvitenskapnb_NO
dc.titleCREEK and Description Logicsnb_NO
dc.typeMaster thesisnb_NO
dc.source.pagenumber114nb_NO
dc.contributor.departmentNorges teknisk-naturvitenskapelige universitet, Fakultet for informasjonsteknologi, matematikk og elektroteknikk, Institutt for datateknikk og informasjonsvitenskapnb_NO


Tilhørende fil(er)

Thumbnail
Thumbnail
Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel