Building a case base for intelligent decision support

Abstract

The support provided to a decision-maker can be strengthened by utilising the past experience represented in a decision support system (DSS) as a memory aid. Past decision situations represented as cases can be used to build the knowledge base component of the case-based decision support system. The cases recording past decisions become a form of organisational memory to support future decision-making. Development of case-based systems requires identification of suitable cases. This is a knowledge acquisition activity where the relevant information is elicited from an expert decision-maker and represented in a case-based decision support system. This paper examines the issue of eliciting past cases and their characteristics from an expert decision-maker. A method is proposed to facilitate the identification of cases and provide a framework for indexing cases in the case-based decision support system. The method employs techniques from knowledge acquisition where the repertory grid approach is used to generate a set of cases and identify their significant characteristics. The degree to which this set of cases is representative of the domain is investigated using a technique for creative and exploratory thought. The method is illustrated with a practical example and its performance is discussed.Presented at: 1st Australian Workshop on Intelligent Decision Support Systems; 1995 Nov 13-17; Canberra, Australia. p. 25-35