|
Abstract : |
Case based reasoning (CBR) uses the knowledge from former experiences ("known cases"). Since special knowledge of an expert is mainly subject to his experiences, the CBR techniques are a good base for the development of expert systems. We investigate the problem for technical diagnosis. Diagnosis is not considered as a classification task, but as a process to be guided by computer assisted experience. This corresponds to the flexible "case completion " approach. Flexibility is also needed for the expert view with predominant interest in the unexpected, unpredictible cases. Integration of further knowledge sources (domain knowledge, common knowledge) is investigated in the project. The system Fallexperte-D is developed as a research and demonstration tool. The flexible composition of cases from basic information entities leads to the use of Case Retrieval Nets (CRN). 1, |
