SURVEY AND CRITIQUE OF TECHNIQUES FOR EXTRACTING RULES FROM TRAINED ARTIFICIAL NEURAL NETWORKS

It is becoming increasingly apparent that, without some form of explan ation capability, the full potential of trained artificial neural netw orks (ANNs) may not be realised. This survey gives an overview of tech niques developed to redress this situation. Specifically, the survey f ocuses on mechanisms, procedures, and algorithms designed to insert kn owledge into ANNs (knowledge initialisation), extract rules from train ed ANNs (rule extraction), and utilise ANNs to refine existing rule ba ses (rule refinement). The survey also introduces a new taxonomy for c lassifying the various techniques, discusses their modus operandi, and delineates criteria for evaluating their efficacy.