论文信息 - PAV and the ROC convex hull

PAV and the ROC convex hull

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.

Tom Fawcett | Alexandru Niculescu-Mizil | Tom Fawcett | Alexandru Niculescu-Mizil

[1] G. Brier. VERIFICATION OF FORECASTS EXPRESSED IN TERMS OF PROBABILITY , 1950 .

[2] Rich Caruana,et al. An empirical comparison of supervised learning algorithms , 2006, ICML.

[3] Tom Fawcett,et al. Robust Classification for Imprecise Environments , 2000, Machine Learning.

[4] J A Swets,et al. Measuring the accuracy of diagnostic systems. , 1988, Science.

[5] J A Swets,et al. Better decisions through science. , 2000, Scientific American.

[6] Peter A. Flach,et al. Repairing Concavities in ROC Curves , 2005, IJCAI.

[7] H. D. Brunk,et al. AN EMPIRICAL DISTRIBUTION FUNCTION FOR SAMPLING WITH INCOMPLETE INFORMATION , 1955 .

[8] John Platt,et al. Probabilistic Outputs for Support vector Machines and Comparisons to Regularized Likelihood Methods , 1999 .

[9] Alexander J. Smola,et al. Advances in Large Margin Classifiers , 2000 .

[10] Bianca Zadrozny,et al. Transforming classifier scores into accurate multiclass probability estimates , 2002, KDD.

[11] W. John Wilbur,et al. The Synergy Between PAV and AdaBoost , 2005, Machine Learning.

[12] Bianca Zadrozny,et al. Learning and making decisions when costs and probabilities are both unknown , 2001, KDD '01.

[13] Rich Caruana,et al. Predicting good probabilities with supervised learning , 2005, ICML.