On the consistency management of large case bases: the case for validation
Kirsti Racine, Qiang Yang · 1996
Case-based reasoning #CBR# is a practical, relatively new technology. CBR is based on the idea that new problems can often be solved by using past solutions. The basic method used to implement CBR is to build a case base of previously solved problems. These cases are then retrieved and adapted to solve new problems. Using this CBR process, a case-based system can learn incrementally and improve its performance over time. However, a pervasive, yet relatively ignored, problem inherent in using this approach is the possible presence of inconsistencies within and among cases. These can be in the form of contradictions within the case base, possibly causing a degradation of performance e#ciency, the retrieval of two con#icting solutions or no retrieval at all. Past research has only dealt with the problem super#cially. In this paper, we present an analysis of inconsistency problems arising from contradiction in a potentially large case base. We classify these problems according to their nature, and suggest validation solutions to deal with them e#ectively.