Tarpeian Bloat Control and Generalization Accuracy

Created by W.Langdon from gp-bibliography.bib Revision:1.4420

  author =       "S{\'e}bastien Mahler and Denis Robilliard and 
                 Cyril Fonlupt",
  editor =       "Maarten Keijzer and Andrea Tettamanzi and 
                 Pierre Collet and Jano I. {van Hemert} and Marco Tomassini",
  title =        "Tarpeian Bloat Control and Generalization Accuracy",
  booktitle =    "Proceedings of the 8th European Conference on Genetic
  publisher =    "Springer",
  series =       "Lecture Notes in Computer Science",
  volume =       "3447",
  year =         "2005",
  address =      "Lausanne, Switzerland",
  month =        "30 " # mar # " - 1 " # apr,
  organisation = "EvoNet",
  keywords =     "genetic algorithms, genetic programming",
  ISBN =         "3-540-25436-6",
  pages =        "203--214",
  DOI =          "doi:10.1007/b107383",
  bibsource =    "DBLP, http://dblp.uni-trier.de",
  abstract =     "In this paper we will focus on machine-learning issues
                 solved with Genetic Programming (GP). Excessive code
                 growth or bloat often happens in GP , greatly slowing
                 down the evolution process. Poli proposed the Tarpeian
                 Control method to reduce bloat, but possible
                 side-effects of this method on the generalisation
                 accuracy of GP hypotheses remained to be tested. In
                 particular, since Tarpeian Control puts a brake on code
                 growth, it could behave as a kind of Occam's razor,
                 promoting shorter hypotheses more able to extend their
                 knowledge to cases apart from any learning steps. To
                 answer this question, we experiment Tarpeian Control
                 with symbolic regression. The results are contrasted,
                 showing that it can either increase or reduce the
                 generalization power of GP hypotheses, depending on the
                 problem at hand. This suggest that a blind use of TC is
                 not safe, but also that a careful parameter setup may
                 be profitable in some cases.",
  notes =        "Part of \cite{keijzer:2005:GP} EuroGP'2005 held in
                 conjunction with EvoCOP2005 and


Genetic Programming entries for Sebastien Mahler Denis Robilliard Cyril Fonlupt