An Evolutionary Approach to Adaptive Model-Building

Learning an input-output mapping from a set of examples can be regarded as a problem of model-building. From this point of view, by a hierarchical representation scheme for models, an evolutionary computational approach to model-building problems is proposed in this paper, which is based on the ideas from the evolution programs and the genetic programmings. The computer experiments indicate that it is surprisingly effective in searching the optimal model for model-building problems.