Timezone: »
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.
Author Information
Stephan Clémençon (Telecom ParisTech)
Nicolas Vayatis (Ecole Normale Supérieure de Cachan)
More from the Same Authors
-
2021 : Handling Distribution Shift in Tire Design »
Antoine De mathelin · François Deheeger · Mathilde MOUGEOT · Nicolas Vayatis -
2022 Poster: What are the best Systems? New Perspectives on NLP Benchmarking »
Pierre Colombo · Nathan Noiry · Ekhine Irurozki · Stephan Clémençon -
2018 Poster: On Binary Classification in Extreme Regions »
Hamid Jalalzai · Stephan Clémençon · Anne Sabourin -
2016 Poster: On Graph Reconstruction via Empirical Risk Minimization: Fast Learning Rates and Scalability »
Guillaume Papa · Aurélien Bellet · Stephan Clémençon -
2014 Poster: Tight Bounds for Influence in Diffusion Networks and Application to Bond Percolation and Epidemiology »
Remi Lemonnier · Kevin Scaman · Nicolas Vayatis -
2012 Poster: Link Prediction in Graphs with Autoregressive Features »
Emile Richard · Stephane Gaiffas · Nicolas Vayatis -
2010 Poster: Link Discovery using Graph Feature Tracking »
Emile Richard · Nicolas Baskiotis · Theos Evgeniou · Nicolas Vayatis -
2009 Poster: AUC optimization and the two-sample problem »
Stéphan Clémençon · Nicolas Vayatis · Marine Depecker -
2008 Poster: Empirical performance maximization for linear rank statistics »
Stephan Clémençon · Nicolas Vayatis -
2008 Poster: On Bootstrapping the ROC Curve »
Patrice Bertail · Stephan Clémençon · Nicolas Vayatis