Extracting decision trees from trained neural networks

In this paper we present a methodology for extracting decision trees from i nput data generated from trained neural networks instead of doing it direct ly from the data. A genetic algorithm is used to query the trained network and extract prototypes. A prototype selection mechanism is then used to sel ect a subset of the prototypes. Finally, a standard induction method like I D3 or C5.0 is used to extract the decision tree. The extracted decision tre es can be used to understand the working of the neural network besides perf orming classification. This method is able to extract different decision tr ees of high accuracy and comprehensibility from the trained neural network. (C) 1999 Pattern Recognition Society. Published by Elsevier Science Ltd. A ll rights reserved.