Accuracy, parsimony, and generality in evolutionary learning systems via multiobjective selection

Xavier Llorà, David E. Goldberg, Ivan Traus, Ester Bernadó

    Producció científica: Capítol de llibreContribució a congrés/conferènciaAvaluat per experts

    17 Cites (Scopus)

    Resum

    Evolutionary learning systems (also known as Pittsburgh learning classifier systems) need to balance accuracy and parsimony for evolving high quality general hypotheses. The learning process used in evolutionary learning systems is based on a set of training instances that sample the target concept to be learned. Thus, the learning process may overfit the learned hypothesis to the given set of training instances. In order to address some of these issues, this paper introduces a multiobjective approach to evolutionary learning systems. Thus, we translate the selection of promising hypotheses into a two-objective problem that looks for: (1) accurate (low error), and (2) compact (low complexity) solutions. Using the proposed multiobjective approach a set of compromise hypotheses are spread along the Pareto front. We also introduce a theory of the impact of noise when sampling the target concept to be learned, as well as the appearance of overfitted hypotheses as the result of perturbations on high quality generalization hypotheses in the Pareto front.

    Idioma originalAnglès
    Títol de la publicacióLearning Classifier Systems - 5th International Workshop, IWLCS 2002, Revised Papers
    EditorsPier Luca Lanzi, Wolfgang Stolzmann, Stewart W. Wilson, Stewart W. Wilson
    EditorSpringer Verlag
    Pàgines118-142
    Nombre de pàgines25
    ISBN (imprès)3540205446, 9783540205449
    DOIs
    Estat de la publicacióPublicada - 2003
    Esdeveniment5th International Workshop on Learning Classifier Systems, IWLCS 2002 - Granada, Spain
    Durada: 7 de set. 20028 de set. 2002

    Sèrie de publicacions

    NomLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volum2661
    ISSN (imprès)0302-9743
    ISSN (electrònic)1611-3349

    Conferència

    Conferència5th International Workshop on Learning Classifier Systems, IWLCS 2002
    País/TerritoriSpain
    CiutatGranada
    Període7/09/028/09/02

    Fingerprint

    Navegar pels temes de recerca de 'Accuracy, parsimony, and generality in evolutionary learning systems via multiobjective selection'. Junts formen un fingerprint únic.

    Com citar-ho