The landscape contest at ICPR 2010

Núria Macià*, Tin Kam Ho, Albert Orriols-Puig, Ester Bernadó-Mansilla

*Autor corresponent d’aquest treball

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

    7 Cites (Scopus)

    Resum

    The landscape contest provides a new and configurable framework to evaluate the robustness of supervised classification techniques and detect their limitations. By means of an evolutionary multiobjective optimization approach, artificial data sets are generated to cover reachable regions in different dimensions of data complexity space. Systematic comparison of a diverse set of classifiers highlights their merits as a function of data complexity. Detailed analysis of their comparative behavior in different regions of the space gives guidance to potential improvements of their performance. In this paper we describe the process of data generation and discuss performances of several well-known classifiers as well as the contestants' classifiers over the obtained data sets.

    Idioma originalAnglès
    Títol de la publicacióRecognizing Patterns in Signals, Speech, Images, and Videos - ICPR 2010 Contests, Contest Reports
    Pàgines29-45
    Nombre de pàgines17
    DOIs
    Estat de la publicacióPublicada - 2010
    Esdeveniment20th International Conference on Pattern Recognition, ICPR 2010 - Istanbul, Turkey
    Durada: 23 d’ag. 201026 d’ag. 2010

    Sèrie de publicacions

    NomLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volum6388 LNCS
    ISSN (imprès)0302-9743
    ISSN (electrònic)1611-3349

    Conferència

    Conferència20th International Conference on Pattern Recognition, ICPR 2010
    País/TerritoriTurkey
    CiutatIstanbul
    Període23/08/1026/08/10

    Fingerprint

    Navegar pels temes de recerca de 'The landscape contest at ICPR 2010'. Junts formen un fingerprint únic.

    Com citar-ho