ISyE SEMINAR SERIES - Inference for an ROC curve
For measuring the accuracy of a continuous diagnostic test, the receiver operating characteristic (ROC) curve is often used. Recently kernel smoothing estimator with two bandwidths for the ROC curve was showed to be better than empirical estimator. In this paper we propose another type of kernel
smoothing estimators which involves only one bandwidth and does not have boundary effect. Moreover, empirical likelihood based confidence interval is obtained.