Deep ROC Analysis and AUC as Balanced Average Accuracy, for Improved Classifier Selection, Audit and Explanation

AMAndre M. CarringtonDGDouglas G. ManuelPWPaul W. FieguthTRTim RamsayVOVenet Osmani

University of Waterloo · Ottawa Hospital · +9 more institutions

PubMed
Indexed incrossrefpubmed

Abstract

Optimal performance is desired for decision-making in any field with binary classifiers and diagnostic tests, however common performance measures lack depth in information. The area under the receiver operating characteristic curve (AUC) and the area under the precision recall curve are too general because they evaluate all decision thresholds including unrealistic ones. Conversely, accuracy, sensitivity, specificity, positive predictive value and the F1 score are too specific-they are measured at a single threshold that is optimal for some instances, but not others, which is not equitable. In between both approaches, we propose deep ROC analysis to measure performance in multiple groups of predicted risk…

Citation impact

279
total citations
FWCI
34.16
Percentile
100%
References
60
Citations per year

Authors

12
  • AM
    Andre M. CarringtonCorresponding

    University of Waterloo, Ottawa Hospital

  • DG
    Douglas G. Manuel

    University of Ottawa, Ottawa Hospital, Institute for Clinical Evaluative Sciences

  • PW
    Paul W. Fieguth

    University of Waterloo

  • TR
    Tim Ramsay

    University of Ottawa, Ottawa Hospital

  • VO
    Venet Osmani

    University of Trento, Fondazione Bruno Kessler

Topics & keywords

Keywords
  • Receiver operating characteristic
  • Precision and recall
  • Binary classification
  • Classifier (UML)
  • False positive rate
  • Pattern recognition (psychology)
  • Binary number
  • Correlation
No related works found for this paper.

Funding