Robust reductions from ranking to classification

M.F. Balcan, N. Bansal, A. Beygelzimer, D. Coppersmith, J. Langford, G.B. Sorkin

    Onderzoeksoutput: Bijdrage aan tijdschriftTijdschriftartikelAcademicpeer review

    37 Citaten (Scopus)

    Samenvatting

    We reduce ranking, as measured by the Area Under the Receiver Operating Characteristic Curve (AUC), to binary classification. The core theorem shows that a binary classification regret of r on the induced binary problem implies an AUC regret of at most 2r. This is a large improvement over approaches such as ordering according to regressed scores, which have a regret transform of r bar right arrow nr where n is the number of elements.
    Originele taal-2Engels
    Pagina's (van-tot)139-153
    TijdschriftMachine Learning
    Volume72
    Nummer van het tijdschrift1-2
    DOI's
    StatusGepubliceerd - 2008

    Citeer dit