Inconsistency - Friend or Foe

Created by W.Langdon from gp-bibliography.bib Revision:1.4504

  author =       "Ulf Johansson and Rikard Konig and Lars Niklasson",
  title =        "Inconsistency - Friend or Foe",
  booktitle =    "International Joint Conference on Neural Networks,
                 IJCNN 2007",
  year =         "2007",
  pages =        "1383--1388",
  address =      "Orlando, USA",
  month =        "12-17 " # aug,
  keywords =     "genetic algorithms, genetic programming, G-REX tree,
                 consistency criterion, evolutionary algorithms,
                 inconsistency criterion, neural network ensembles,
                 probability estimation, publicly available data sets,
                 regression trees, rule extraction algorithms, data
                 integrity, data mining, estimation theory, evolutionary
                 computation, learning (artificial intelligence),
                 probability, regression analysis",
  ISSN =         "1098-7576",
  isbn13 =       "1-4244-1380-X",
  DOI =          "doi:10.1109/IJCNN.2007.4371160",
  abstract =     "One way of obtaining accurate yet comprehensible
                 models is to extract rules from opaque predictive
                 models. When evaluating rule extraction algorithms, one
                 frequently used criterion is consistency; i.e. the
                 algorithm must produce similar rules every time it is
                 applied to the same problem. Rule extraction algorithms
                 based on evolutionary algorithms are, however,
                 inherently inconsistent, something that is regarded as
                 their main drawback. In this paper, we argue that
                 consistency is an over valued criterion, and that
                 inconsistency can even be beneficial in some
                 situations. The study contains two experiments, both
                 using publicly available data sets, where rules are
                 extracted from neural network ensembles. In the first
                 experiment, it is shown that it is normally possible to
                 extract several different rule sets from an opaque
                 model, all having high and similar accuracy. The
                 implication is that consistency in that perspective is
                 useless; why should one specific rule set be considered
                 superior? Clearly, it should instead be regarded as an
                 advantage to obtain several accurate and comprehensible
                 descriptions of the relationship. In the second
                 experiment, rule extraction is used for probability
                 estimation. More specifically, an ensemble of extracted
                 trees is used in order to obtain probability estimates.
                 Here, it is exactly the inconsistency of the rule
                 extraction algorithm that makes the suggested approach
  notes =        "Also known as \cite{4371160}",

Genetic Programming entries for Ulf Johansson Rikard Konig Lars Niklasson