Introduction and overview of case-based reasoning

Case-based reasoning: Foundational issues, methodological variations and system approaches

An introduction to case-based reasoning and a framework for describing CBR systems. The paper introduces the 'CBR cycle', which has been adopted as a standard reference for the basic CBR method.

Published in AI Communications - The European Journal of Artificial Intelligence, Vol. 7, no. 1 (1994), pp. 39-59. Copyright: IOS Press.

Different roles and mutual dependencies of data, information and knowledge

Here we look at the CBR method within the context of information systems and database systems. A definition of core concepts is given, and an approach is presented that can take a system from a database system, through an information system, to a konledge-based system exemplified by a CBR system.

Published in Data and Knowledge Engineering, Vol. 16 (1995), pp. 191-222. Copyright: Elsevier.

Case-based reasoning foundations

A brief summary of the historical development of CBR technology with expamples of some early systems.

Published in The Knowledge Engineering Review, Vol. 20, no. 3 (2006), pp. 203-207. Copyright: Cambridge University Press.

Retrieval, re-use, revision and retention in case-based reasoning

A comprehensive paper that presents ad contrasts a variety of CBR methods, with a reference to the four-step CBR cycle. Can be viewed as a thorough update of the AI Communications paper from 1994.

Published in The Knowledge Engineering Review, Vol. 20, no. 3 (2006), pp. 215-240. Copyright: Cambridge University Press.

Explanation-driven case-based reasoning

An early paper that presents the CBR system Creek, which is the originator if Verdande Technology's CBR systems. An emphasis is put on how general domain knowledge is used as support knowledge in the case-based reasoning process.

Published in Lecture Notes in Computer Science, Topics in Case-Based Reasoning, Vol. 837/1994, pp. 274-288. Copyright: Springer.

Knowledge-intensive case-based reasoning in Creek

This paper introduces CBR as a knowledge modelling and problem solving methodology, and continues to describe the Troll Creek system, a Java implementation which supports the construction and execution of Creek knowledge bases. Examples of PhD research projects within TrollCreek, as well as other projects, are given. TrollCreek was the version of Creek that formed the direct basis for out DrillEdge software.

Published in 6th International Conference on Case-Based Reasoning, Workshop Proceedings. DePaul University, Chicago, 2005. pp.62-71.

Modelling of cases and associated general knowledge

Knowledge acquisition and learning from experience - the role of case-spesific knowledge

This is a rather long paper that covers many aspects of knowledge acquisition, knowledge modelling and machine learning. Its focus is the CBR-view to these tasks. A framework is suggested that combines initial and periodic top-down acquisition and modelling of knowledge with bottom-up learning within the model based on problem solving experience. Examples from Creek are included.

Published in Gheorge Tecuci and Yves Kodratoff (eds): Machine learning and knowledge acquisition; Integrated approaches, (1995), pp. 197-245. Copyright: Academic Press.

Integrating Bayesian networks into knowledge-intensive CBR

A method is described that combines CBR with a probabilistic causal model (i.e. a Bayesian Network). The BN represents general domain knowledge as cause-effect relations, and is used as a knowledge-based method to improve the search for a best matching case. An example from the car starting domain illustrates the method.

Modelling the knowledge contents of CBR systems

Here, the role of the 'knowledge level' a conceptual modelling level for knowledge-based systems, in building CBR systems is briefly discussed. An example, using the 'Components of Expertise' knowledge level methodology is given. It is a position paper, which tries to motivate the use of the knowledge level in CBR.

Published in Proceedings of the Workshop Program at the Fourth International Conference on Case-Based Reasoning, Vancouver, 2001. Naval Research Laboratory Technical Note AIC-01-003, pp. 32-37.

Cases for providing explanations

Explanation in case-based reasoning, perspectives and goals

This is a comprehensive survey on the role of CBR for generating explanations that help a user in assessing the advice coming from a system. A set of explanation goals such transparency and justifiability are defined, and relevant CBR roles and methods are pointed to and discussed.

Published in Artificial Intelligence Review. Vol 24, no. 2, October 2005. pp. 109-143. Copyright: Springer.

Explanatory capabilities in the CREEK knowledge-intensive case-based reasoner 

This paper discussed the capabilities for producing useful explanations within the Creek system, with reference to the explanation goals defined in the above paper. An example is given using an ambient intelligence implementation of Creek, where the CBR engine is implemented in an agent architecture. An example of its use for situation recognition in a hospital ward is included.

Published in Frontiers in Artificial Intelligence and Applications, Vol. 172 (2008), Tenth Scandinavian Conference on Arificial Intelligence, pp. 28-35. Copyright: IOS Press.