• Asia pacific journal of information systems
• /
• v.14 no.3
• /
• pp.77-92
• /
• 2004

When humans make decisions, they differentiate classifications of individual attribute variables that affect the decisions according to the importance and pattern of each attribute variables. The present study examines the practicality of the proposed Code Arrangement-Based Reasoning(CABR), which resembles the human's way of reasoning. To this end, we developed a CABR technique that classifies each attribute variable affecting significant impacts on the target variable into a cluster and assigns a code to the cluster. For verifying the proposed technique, both case-based reasoning and CABR were used for the customer continuance judgment problem of an automobile insurance company. Results indicated that the performance of CABR is close to the one of the case-based reasoning. The CABR also shows the possibility of using bio-informatics techniques for organizational data analysis in the future.