Paper 80: "Robust Adaptation in CBR"


Case adaptation largely determines the flexibility of a Case-based Reasoning (CBR) system by potentially providing unseen solutions adapted from stored knowledge. Traditional case adaptation is done a single-step process, attempting to modify the retrieved case to be closer to the query at hand by certain adaptation rules. A CBR system can modifying a retrieved case multiple times using multiple rules, resulting in a chain of adapted cases and adaptation rules, or an adaptation path. This research proposes techniques to enable to robust adaptation (ROAD) for the building of adaptation paths. Moreover, the adaptation paths built offers introspective learning opportunities to improve aspects of the CBR system other than the case adaptation process.

0 commentaries