Estimating weighted areas under the ROC curve

Andreas Maurer; Massimiliano Pontil

NeurIPS 2020poster7 citations

Estimating weighted areas under the ROC curve

Abstract

Exponential bounds on the estimation error are given for the plug-in estimator of weighted areas under the ROC curve. The bounds hold for single score functions and uniformly over classes of functions, whose complexity can be controlled by Gaussian or Rademacher averages. The results justify learning algorithms which select score functions to maximize the empirical partial area under the curve (pAUC). They also illustrate the use of some recent advances in the theory of nonlinear empirical processes.

BibTeX

@inproceedings{NEURIPS2020_5781a263,
 author = {Maurer, Andreas and Pontil, Massimiliano},
 booktitle = {Advances in Neural Information Processing Systems},
 editor = {H. Larochelle and M. Ranzato and R. Hadsell and M.F. Balcan and H. Lin},
 pages = {7733--7742},
 publisher = {Curran Associates, Inc.},
 title = {Estimating weighted areas under the ROC curve},
 url = {https://proceedings.neurips.cc/paper_files/paper/2020/file/5781a2637b476d781eb3134581b32044-Paper.pdf},
 volume = {33},
 year = {2020}
}