Generating Summaries from Retrieved Base Cases

Despite ongoing efforts a lack of efficient case evaluation remains a central problem within the relatively young research area of case-based reasoning (CBR). In simple terms case evaluation is the task of identifying the problem-solving potential retrieved cases provide for a given situation. Very frequently this process includes the identification and generation of summaries for all, or some retrieved candidate cases. In many applications case evaluation is done by the system user. This paper addresses the problem by presenting a general method for case evaluation. The method generates expressive summaries for retrieved cases based on the theory of fuzziness. Results from an application in the domain of Coronary Heart Disease Risk Assessment (CHDRA) indicate the value of the method for case evaluation. The paper also identifies the potential of the method for other CBR issues like similarity assessment and case indexing.