Increasing Rule Extraction Accuracy by Post-processing GP Trees
2008 (English)In: Proceedings of the Congress on Evolutionary Computation, IEEE Press , 2008, 3010-3015 p.Conference paper (Refereed)
Genetic programming (GP), is a very general and efficient technique, often capable of outperforming more specialized techniques on a variety of tasks. In this paper, we suggest a straightforward novel algorithm for post-processing of GP classification trees. The algorithm iteratively, one node at a time, searches for possible modifications that would result in higher accuracy. More specifically, the algorithm for each split evaluates every possible constant value and chooses the best. With this design, the post-processing algorithm can only increase training accuracy, never decrease it. In this study, we apply the suggested algorithm to GP trees, extracted from neural network ensembles. Experimentation, using 22 UCI datasets, shows that the post-processing results in higher test set accuracies on a large majority of datasets. As a matter of fact, for two setups of three evaluated, the increase in accuracy is statistically significant.
Place, publisher, year, edition, pages
IEEE Press , 2008. 3010-3015 p.
genetic programming, rule extraction, Computer Science, Machine Learning, Data Mining
Computer and Information Science
IdentifiersURN: urn:nbn:se:hb:diva-5947Local ID: 2320/3975ISBN: 978-1-4244-1823-7OAI: oai:DiVA.org:hb-5947DiVA: diva2:886630
CEC 2008, Hong Kong, June 1- 6, 2008
This work was supported by the Information Fusion Research Program (University of Skövde, Sweden) in partnership with the Swedish Knowledge Foundation under grant 2003/0104 (URL: http://www.infofusion.se).2015-12-222015-12-22