Mukesh Patel, Uwe Schnepf
In this paper we describe the kind of learning phenomena we are interested in, i.e. concept formation, and we discuss some of the shortcomings of traditional knowledge-based systems. After that, our approach will be introduced, first as a theory and then as a description. Following which, induction as the main constituent of learning capabilities is described. We describe the learning mechanisms that we use to investigate the problem. This in our case are Classifier Systems and Genetic Algorithms. We roughly explain what they are, how they work and what sorts of abstractions they perform. Finally, a scenario will be presented which serves to validate the approach chosen.
This paper is not available online