Relevance rule based reasoning

Case based reasoning and rule based deductive reasoning have long been considered as two different reasoning paradigms. To integrate the two, this paper first introduce a new form of knowledge, the relevance rule (R-rule), which represents knowledge about relevant relationship between attributes. The R-rule based reasoning paradigm is then described. The inductive learning of the R-rule is also briefly addresses. In the end, it is argued that the R-rule enables the unification of case base reasoning and rule based reasoning in a single reasoning paradigm.