A0276
Title: Bayesian-frequentist hybrid inference in clinical and genomic research applications
Authors: Gang Han - Texas A&M University (United States) [presenting]
Abstract: The Bayesian-frequentist hybrid model and associated inference can combine the advantages of both Bayesian and frequentist methods and avoid their limitations. However, except for a few special cases in existing literature, the computation under the hybrid model is generally nontrivial or even unsolvable. A computation algorithm for hybrid inference is developed under any general loss functions. Three simulation examples demonstrate that hybrid inference can improve upon frequentist inference by incorporating valuable prior information. Bayesian inference based on non-informative priors is also improved, where the latter leads to biased estimates for the small sample sizes used in inference. The proposed method is illustrated in research applications, including a biomechanical engineering design for knee prostheses, surgical treatment of acral lentiginous melanoma, modelling HIV viral load dynamics, and an analysis of RNA single-cell sequencing data incorporating cell probabilities for identifying protein-coding genes associated with pulmonary fibrosis.
