Structural cases representation in case-based expert systems

Abstract

In recent years, a lot of case-based expert systems (CBES) have been proposed to solve problems in many domains, bur how to effectively retrieve approximate case is still a bottleneck in building CBES. In this work, the embedded relationships between the cases are used to solve the above problem. To elicit the knowledge embedded in the case structure, a new knowledge representation named Structural Cases (SC) which is the basis of our system, is proposed to describe the relationships between two different cases. Based upon this representation, a process of transforming the cases into a set of structural cases is proposed, where two cases with similarity greater than the given threshold are merged into one structural case. Moreover the algorithms for interaction process based on the Structural Cases is also proposed which can be used to retrieve more suitable results. Finally, some examples are given to demonstrate the practicality of the structural cases.