ML Anthology

PAV and the ROC Convex Hull

Tom Fawcett, Alexandru Niculescu-Mizil
MLJ 2007 pp. 97-106
doi:10.1007/S10994-007-5011-0 /mlj/2007/fawcett2007mlj-pav/

Abstract

Classifier calibration is the process of converting classifier scores into reliable probability estimates. Recently, a calibration technique based on isotonic regression has gained attention within machine learning as a flexible and effective way to calibrate classifiers. We show that, surprisingly, isotonic regression based calibration using the Pool Adjacent Violators algorithm is equivalent to the ROC convex hull method.

PDF MLJ Semantic Scholar

Cite

Text

Fawcett and Niculescu-Mizil. "PAV and the ROC Convex Hull." Machine Learning, 2007. doi:10.1007/S10994-007-5011-0

Markdown

[Fawcett and Niculescu-Mizil. "PAV and the ROC Convex Hull." Machine Learning, 2007.](https://mlanthology.org/mlj/2007/fawcett2007mlj-pav/) doi:10.1007/S10994-007-5011-0

BibTeX

@article{fawcett2007mlj-pav,
  title     = {{PAV and the ROC Convex Hull}},
  author    = {Fawcett, Tom and Niculescu-Mizil, Alexandru},
  journal   = {Machine Learning},
  year      = {2007},
  pages     = {97-106},
  doi       = {10.1007/S10994-007-5011-0},
  volume    = {68},
  url       = {https://mlanthology.org/mlj/2007/fawcett2007mlj-pav/}
}