Title :
Bayesian confidence intervals for ROC curves
Author :
Parker, D.R. ; Gustafson, S.C. ; Ross, T.D.
Author_Institution :
Dept. of Electr. Eng.& Comput. Sci., Air Force Inst. of Technol., Wright-Patterson AFB, OH, USA
fDate :
3/3/2005 12:00:00 AM
Abstract :
The probability density of correct detection probability for any false alarm probability on a receiver operating characteristic (ROC) curve is obtained using Bayesian methods. Simulations demonstrate that this density provides rigorous and practical methods for obtaining ROC curve confidence intervals.
Keywords :
Bayes methods; probability; sensitivity analysis; signal detection; Bayesian confidence intervals; ROC curve confidence intervals; correct detection probability; false alarm probability; probability density; receiver operating characteristic curve;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20047523
