search for




 

Three-way partial VUS
Journal of the Korean Data & Information Science Society 2019;30:445-54
Published online March 31, 2019;  https://doi.org/10.7465/jkdi.2019.30.2.445
© 2019 Korean Data and Information Science Society.

Chong Sun Hong1  Min Sub Jung2  Hye Soo Shin3

123Department of Statistics, Sungkyunkwan University
Correspondence to: Professor, Department of Statistics, Sungkyunkwan University, Seoul 03063, Korea. E-mail: cshong@skku.edu
Received February 13, 2019; Revised March 16, 2019; Accepted March 16, 2019.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
The AUC and VUS for the ROC curve and surface are well known measures in various statistical elds. Since the partial AUC has only conditions on the FPR, this overlapped condition under the constraints of the TPR may reduce the eciency and accuracy of the performance evaluation. For obtaining better performance measurement for high TPR and low FPR, the two-way partial AUC was proposed. In this work, the partial AUC and two-way partial AUC are de ned using the probability concept, and these are expressed with integration equations. Moreover, the partial VUS and three- way partial VUS for ROC surface are also derived.
Keywords : Classi cation, FPR, sensitivity, speci city, TPR.