Overlaying classifiers: a practical approach for optimal ranking

Abstract : ROC curves are one of the most widely used displays to evaluate performance of scoring functions. In the paper, we propose a statistical method for directly optimizing the ROC curve. The target is known to be the regression function up to an increasing transformation and this boils down to recovering the level sets of the latter. We propose to use classifiers obtained by empirical risk minimization of a weighted classification error and then to construct a scoring rule by overlaying these classifiers. We show the consistency and rate of convergence to the optimal ROC curve of this procedure in terms of supremum norm and also, as a byproduct of the analysis, we derive an empirical estimate of the optimal ROC curve.
Complete list of metadatas

Cited literature [16 references]  Display  Hide  Download

https://hal.telecom-paristech.fr/hal-02107204
Contributor : Stephan Clémençon <>
Submitted on : Tuesday, April 23, 2019 - 3:35:04 PM
Last modification on : Wednesday, July 3, 2019 - 3:02:03 PM

File

3469-overlaying-classifiers-a-...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02107204, version 1

Citation

Stéphan Clémençon, Nicolas Vayatis. Overlaying classifiers: a practical approach for optimal ranking. Overlaying classifiers: a practical approach for optimal ranking, 2009, Advances in Neural Information Processing Systems 21. ⟨hal-02107204⟩

Share

Metrics

Record views

11

Files downloads

6