IMPLEMENTATION AND REFINEMENT OF DECISION TREES USING NEURAL NETWORKSFOR HYBRID KNOWLEDGE ACQUISITION

Decision tree induction is one of the most effective techniques for ac quiring classification knowledge. However, appropriate pre- and post-p rocessors have to be prepared to achieve continuous input/output mappi ng, because the decision trees basically deal with symbolic knowledge. On the other hand, an artificial neural network is suitable for such a purpose, however, its initial structure is difficult to constitute. The authors' research goal is to develop a sophisticated knowledge acq uisition system integrating decision tree induction for identifying th e fundamental structure of the knowledge and neural network generation for realizing an adaptive processor based on the knowledge structure obtained as a decision tree. This paper reports an experimental approa ch to this goal by constructing a neural network based on the result o f decision tree induction from symbolic examples, and analysing the ne twork to elicit hidden knowledge in numerical examples.