Term of Award
Doctor of Public Health in Biostatistics (Dr.P.H.)
Document Type and Release Option
Dissertation (open access)
Copyright Statement / License for Reuse
This work is licensed under a Creative Commons Attribution 4.0 License.
College of Public Health
Committee Member 1
Committee Member 2
Committee Member 3
Committee Member 3 Email
The Kullback-Leibler divergence (KL), which captures the disparity between two distributions, has been considered as a measure for determining the diagnostic performance of an ordinal diagnostic test. This study applies KL and further generalizes it to comprehensively measure the diagnostic accuracy test for multi-stage (K > 2) diseases, named generalized total Kullback-Leibler divergence (GTKL). Also, GTKL is proposed as an optimal cut-points selection criterion for discriminating subjects among different disease stages. Moreover, the study investigates a variety of applications of GTKL on measuring the rule-in/out potentials in the single-stage and multi-stage levels. Intensive simulation studies are conducted to compare the performance of GTKL and other diagnostic accuracy measures, such as generalized Youden index (GYI), hypervolume under the manifold (HUM), and maximum absolute determinant (MADET). Furthermore, a comprehensive analysis of a real dataset is performed to illustrate the application of the proposed measure.
Mo, Chen, "Generalization of Kullback-Leibler Divergence for Multi-Stage Diseases: Application to Diagnostic Test Accuracy and Optimal Cut-Points Selection Criterion" (2020). Electronic Theses and Dissertations. 2046.
Research Data and Supplementary Material
Available for download on Monday, April 05, 2021