Kernel Estimates of ROC Curves

Warning

This publication doesn't include Institute of Computer Science. It includes Faculty of Science. Official publication website can be found on muni.cz.

Authors	ZELINKA Jiří
Year of publication	2008
Type	Article in Proceedings
Conference	Compstat 2008, Books of Abstracts
MU Faculty or unit	Faculty of Science
Citation
Field	Applied statistics, operation research
Keywords	kernel estimate; ROC curve; bandwidth
Description	ROC curves are important tool for comparing diagnostic tests. Nonparametric kernel estimation is useful in the case when no information about the distribution of the data is available. Presented contribution is focused on the method of kernel estimate of ROC curve as a special type of cumulative distribution function. This approach can be applied for choice of smoothing parameter called bandwidth and also for evaluation of area under ROC curve (AUC) or partial AUC. Suggested method is compared with another methods and it is tested on simulated and real data.
Related projects:	Jaroslav Hájek Center for Theoretical and Applied Statistics