Skip to Main Content
The conventional machine learning approaches can provide well established experimental basis for genetic based learners. Examination of the research imply that the link between evolutionary learning and conventional learning studies may be improved. This is especially true in terms of practices adopting generalisation as a performance evaluation criterion for learning. In this paper an overview of significant number of experiments from classifier systems and genetic programming are presented. Suggestions to accommodate generalisation in the context of evolutionary learning are provided within a generic learning framework and its implication for evolutionary generalisation is discussed.