Managing Multiple Case Bases: Dimensions and Issues. (pdf )

David B. Leake and Raja Sooriamurthi. Proceedings of the Fifteenth International Florida Artificial Intelligence Research Society Conference. AAAI Press, Menlo Park, 2001, pp. 106-110. 5 pages.


Case-based reasoning (CBR) models the process of reasoning from specific experiences acquired by an agent, and contained in the agent's case-base. When multiple agents acquire cases, opportunities arise for sharing their case-bases, with accompanying issues for how to apply others' experiences effectively. This paper examines issues for multi-case-base reasoning (MCBR), the reasoning process needed for a CBR system to exploit external case-bases reflecting similar but different tasks and task environments. The paper summarizes the component processes required, the dimensions along which these processes may differ, and some of the key research issues that must be addressed for successful MCBR systems. It closes with a perspective on the relationships of case-based reasoning and multi-case-base reasoning, examining the analogy between reasoning about cases in CBR and case-bases in MCBR.

See for additional publications in the Artificial Intelligence/Cognitive Science report and reprint archive maintained by David Leake.