CASE-BASED REASONING FOR KNOWLEDGE ACQUISITION SUGGESTIONS
Case-based suggestion (CBS) is a general mechanism for system-driven interactive knowledge acquisition. CBS applies case-based reasoning to the task of knowledge acquisition. It utilizes previously acquired knowledge embodied in cases to assist the expert during the current knowledge acquisition session. In this work we describe the general CBS technique and illustrate its use during the acquisition of a specific kind of knowledge. A system utilizing CBS was implemented in the acquisition module of a prototype system called ODS, which structures acquired diagnostic knowledge in decision trees. The algorithm used for case-based suggestion by the ODS system and a description of how the decision tree knowledge was represented in the case base is presented. Several evaluation metrics are introduced, and the application of these measures to several experiences of acquiring knowledge with ODS is discussed.