As the computing world moves from the information age into the knowledge-based age, it is beneficial to induce
knowledge from the information superhighway formed from the Internet and intranet. The knowledge acquired can be
expressed in different knowledge representations such as computer programs, first-order logical relations, or fuzzy Petri nets
Ž . FPNs . In this paper, we present a flexible knowledge discovery system called generic genetic programming GGP that Ž .
applies genetic programming GP and logic grammars to learn knowledge in various knowledge representation formalisms. Ž .
An experiment is performed to demonstrate that GGP can discover knowledge represented in FPNs that support fuzzy and
approximate reasoning. To evaluate the performance of GGP in producing good FPNs, the classification accuracy of the
FPN induced by GGP and that of the decision tree generated by C4.5 are compared. Moreover, the performance of GGP in
inducing logic programs from noisy examples is evaluated. A detailed comparison to FOIL, a system that induces logic
programs, has been conducted. These experiments demonstrate that GGP is a promising alternative to other knowledge
discovery systems and sometimes is superior for handling noisy and inexact data.