Generalized Analytic Rule Extraction for Feedforward Neural Networks

This paper suggests the "lnpUt-NetwOik-Training-0ulput-Extract~~Knowledge" framework to classify existing rule exlraction algorithms for feedloward neural networks. Based on the Suggested framework, we identify the major practices of existing algorithms as relying on the technique Of generate and test, which leads to exponential complexity, relying on specialized network structure and training algorithms, which leads to limited applications and rsliance on the interpretation ot hidden nodes, which leads to proliferation of classification ruies and their incomprehensibility. in order to generalize the applicability of rule extraction. we propose the rule extraction algorithm GeneraLized Analytic Rule Extraction (GLARE), and demonstrate its efficaoy by comparing it with neural networks per se and the popular rule extraction program for decision trees, C4.5. index Term-Ciassiticatian, neural network, rule extraction. +