A Comparison of Linear Genetic Programming and Neural Networks in Medical Data Mining

Created by W.Langdon from gp-bibliography.bib Revision:1.3949

@Article{Brameier:2001:TEC,
  author =       "Markus Brameier and Wolfgang Banzhaf",
  title =        "A Comparison of Linear Genetic Programming and Neural
                 Networks in Medical Data Mining",
  journal =      "IEEE Transactions on Evolutionary Computation",
  year =         "2001",
  volume =       "5",
  number =       "1",
  pages =        "17--26",
  month =        feb,
  keywords =     "genetic algorithms, genetic programming, Data mining,
                 evolutionary computation, neural networks",
  URL =          "http://web.cs.mun.ca/~banzhaf/papers/ieee_taec.pdf",
  size =         "10 pages",
  abstract =     "We apply linear genetic programming to several
                 diagnosis problems in medicine. An efficient algorithm
                 is presented that eliminates intron code in linear
                 genetic programs. This results in a significant speedup
                 which is especially interesting when operating with
                 complex datasets as they are occuring in real-world
                 applications like medicine. We compare our results to
                 those obtained with neural networks and argue that
                 genetic programming is able to show similar performance
                 in classification and generalization even within a
                 relatively small number of generations.",
  notes =        "proben1/UCI LGP variable length string of C
                 instruction. Branching. steady state tournament
                 selection. two-point string crossover {"}high mutation
                 rates have been experienced to produced better
                 results{"} p19. Size<=256 {"}it is much easier for the
                 GP system to implement structural introns [than
                 semantic ones]{"} p20 {"}for all problems discussed,
                 the performance of GP in generalization comes close to
                 or even better then the results documented for NNs{"}
                 (MLP, RPROP) p21 Ten demes of 500 connected in one
                 direction circle. 5% mutation rate. {"}On average, the
                 number of effective generations is reduced by a factor
                 of three when using demes. Tests with and without
                 conditionals. Runtime comparison.

                 Intron removal (dead code) at run time.",
}

Genetic Programming entries for Markus Brameier Wolfgang Banzhaf

Citations