Efficient AUC optimization for classification

T. Calders, S. Jaroszewicz

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    85 Citations (Scopus)
    464 Downloads (Pure)


    In this paper we show an efficient method for inducing classifiers that directly optimize the area under the ROC curve. Recently, AUC gained importance in the classification community as a mean to compare the performance of classifiers. Because most classification methods do not optimize this measure directly, several classification learning methods are emerging that directly optimize the AUC. These methods, however, require many costly computations of the AUC, and hence, do not scale well to large datasets. In this paper, we develop a method to increase the efficiency of computing AUC based on a polynomial approximation of the AUC. As a proof of concept, the approximation is plugged into the construction of a scalable linear classifier that directly optimizes AUC using a gradient descent method. Experiments on real-life datasets show a high accuracy and efficiency of the polynomial approximation.
    Original languageEnglish
    Title of host publicationProceedings of the 11th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD 2007) 17-21 September 2007, Warsaw, Poland
    EditorsJ.N. Kok, J. Koronacki, R. Lopez de Mantaras, S. Matwin, D. Mladenic, A. Skowron
    Place of PublicationBerlin, Germany
    ISBN (Print)978-3-540-74975-2
    Publication statusPublished - 2007
    Eventconference; PKDD 2007, Warsaw, Poland; 2007-09-17; 2007-09-21 -
    Duration: 17 Sept 200721 Sept 2007

    Publication series

    NameLecture Notes in Computer Science
    ISSN (Print)0302-9743


    Conferenceconference; PKDD 2007, Warsaw, Poland; 2007-09-17; 2007-09-21
    OtherPKDD 2007, Warsaw, Poland


    Dive into the research topics of 'Efficient AUC optimization for classification'. Together they form a unique fingerprint.

    Cite this