Case-Based Reasoning (CBR) suffers, like the majority of systems, from a large storage requirement and a slow query execution time, especially when dealing with a large case base. As a result, there has been a significant increase in the research area of Case Base Maintenance (CBM). This paper proposes a case-base maintenance method based on the machine-learning techniques, it is able to maintain the case bases by reducing its size and preserving maximum competence of the system. The main purpose of our method is to apply clustering analysis to a large case base and efficiently build natural clusters of cases which are smaller in size and can easily use simpler maintenance operations. For each cluster we reduce as much as possible, the size of the cluster.
展开▼
