Publication Date
2026
Document Type
Dissertation/Thesis
First Advisor
Ryu, Duchwan
Degree Name
Ph.D. (Doctor of Philosophy)
Legacy Department
Department of Mathematical Sciences
Abstract
Heterogeneity in survival data presents challenges for identifying underlying mixture structures. The α-mixture family offers a flexible framework for capturing diverse survival patterns by balancing failure rates and distributional shapes across heterogeneous populations. Despite its theoretical potential, Bayesian inference for α-mixtures remains underexplored. In this work, we propose a Bayesian approach for estimating α-mixture survival models and evaluating their goodness of fit. We assess the method through four simulation studies involving different mixture combinations, demonstrating robust and reliable parameter recovery across a range of scenarios. To facilitate practical application, we introduce the R package alpmixBayes, which performs Bayesian estimation, provides 95% credible intervals, and leverages the posterior distribution of α to guide mixture selection. Beyond estimation, we apply the framework to real-world datasets to identify the most appropriate mixture types, integrating theoretical considerations with empirical insights. We also extend the model to a more general u-mixture formulation, capable of accommodating multiple mixture types simultaneously. This generalization enhances flexibility in capturing complex population heterogeneity and informs strategies for model interpretation, inference, and practical implementation in reliability and biomedical studies.
Recommended Citation
Luan, Feng, "α-Mixture Survival Model: Estimation and Generalization" (2026). Graduate Research Theses & Dissertations. 8223.
https://huskiecommons.lib.niu.edu/allgraduate-thesesdissertations/8223
Extent
75 pages
Language
en
Publisher
Northern Illinois University
Rights Statement
In Copyright
Rights Statement 2
NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.
Media Type
Text
