Abstract
We demonstrate the use of an unsupervised learning technique called genetic algorithms to discover the association between a concept and its key attributes in concept characterization. The resulting concept- attribute associations are important domain concepts for knowledge engineers to structure interviews with the experts or to prepare representative data for inductive inference. Examples based on the part family identification problem in manufacturing are employed to illustrate the identification capability of our technique. Preliminary results from testing the technique in a SUN SPARC station 1+ indicate that it can be exploited as a decision support tool to assist knowledge engineers in the conceptualization stage of the knowledge acquisition process.
Recommended Citation
Lee-Post, Anita, "Decision Support in Knowledge Acquisition: Concept Characterization Using Genetic Algorithms" (1996). ICIS 1996 Proceedings. 12.
https://aisel.aisnet.org/icis1996/12