Using nomination sampling in estimating the area under the ROC curve.

The area under a receiver operating characteristic (ROC) curve is frequently used in medical studies to evaluate the effectiveness of a continuous diagnostic biomarker, with values closer to one indicating better classification. Unfortunately, the standard statistical procedures based on simple random sampling (SRS) and ranked set sampling (RSS) techniques tend to be less efficient when the values of the area under a ROC curve (AUC) get closer to one. Thus, developing some statistical procedures for efficiently estimating the AUC when it is close to one is very important. In this paper, some estimators are developed using nomination sampling to assess AUC. The proposed AUC estimators are compared with their counterparts in SRS and RSS using Monte Carlo simulation. The results show that some of the estimators developed in this study considerably improve the efficiency of the AUC estimation when it is close to one. This substantially reduces the cost and time for the sample size needed to obtain the desired precision. [ABSTRACT FROM AUTHOR]

Copyright of Computational Statistics is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)