Using Imaginary Ensembles to Select GP Classifiers
2010 (English)In: Genetic Programming: 13th European Conference, EuroGP 2010, Istanbul, Turkey, April 7-9, 2010. Proceedings / [ed] Anna Isabel Esparcia-Alcázar, Anikó Ekárt, Sara Silva, Stephen Dignum, A. Şima Uyar, Springer Berlin/Heidelberg, 2010, p. 278-288Conference paper, Published paper (Refereed)
Abstract [en]
When predictive modeling requires comprehensible models, most data miners will use specialized techniques producing rule sets or decision trees. This study, however, shows that genetically evolved decision trees may very well outperform the more specialized techniques. The proposed approach evolves a number of decision trees and then uses one of several suggested selection strategies to pick one specific tree from that pool. The inherent inconsistency of evolution makes it possible to evolve each tree using all data, and still obtain somewhat different models. The main idea is to use these quite accurate and slightly diverse trees to form an imaginary ensemble, which is then used as a guide when selecting one specific tree. Simply put, the tree classifying the largest number of instances identically to the ensemble is chosen. In the experimentation, using 25 UCI data sets, two selection strategies obtained significantly higher accuracy than the standard rule inducer J48.
Place, publisher, year, edition, pages
Springer Berlin/Heidelberg, 2010. p. 278-288
Series
Lecture Notes in Computer Science (LNCS), ISSN 0302-9743 ; 6021
Keywords [en]
Classification, Decision trees, Genetic programming, Ensembles
National Category
Computer and Information Sciences
Research subject
Technology
Identifiers
URN: urn:nbn:se:his:diva-4020DOI: 10.1007/978-3-642-12148-7_24ISI: 000278827300024Scopus ID: 2-s2.0-77952301837ISBN: 978-3-642-12147-0 ISBN: 978-3-642-12148-7 ISBN: 3-642-12147-0 OAI: oai:DiVA.org:his-4020DiVA, id: diva2:322551
Conference
13th European Conference on Genetic Programming (EuroGP), Istanbul, Turkey, April 7-9, 2010
2010-06-072010-06-072018-01-12Bibliographically approved