M. Egmont-Petersen. "Homomorphic transformation from neural networks
to rule bases," In: E. Mosekilde (Eds.), Proceedings of the european
simulation multiconference, pp. 260-265, 1991.
In this article a method to extract the knowledge induced in a neural network is presented. The method explicates the relation between a network's inputs and its outputs. This relation is stored as logic rules. The feasibility of the method is studied by means of three test examples. The result is that the method can be used, though some drawbacks are detected. One is that the method sometimes generates a lot of rules. For fast retrieval, these rules can well be stored in a B-tree.
Electronic reprint, or contact me: email@example.com