Research on Ontology-Based Case Indexing in CBR

This paper presents methods that support case retrieval in case-based reasoning system. We used the ontology to describe the relationship between terms in application fields. The similar cases are retrieval by calculating semantic similarity which we have defined. We evaluated traditional method of calculating the semantic similarity with lattice theory. We have constructed a decision support CBR prototype system of marketing strategy, based on this algorithm, which contains more than 600 cases. The evaluation shows that with the support of semantic, we can not only carry out data matching retrieval, but also perform semantic associated data access. CBR can quickly and accurately retrieve cases and improve efficiency of reasoning by semantic query.