Efficient Bayesian Methods for Biostatistics
Sampling-based methods like MCMC/HMC is considered the gold standard for Bayesian inference. However, for large data and complex models, they suffer from severe computational cost and issues with convergence. Approximate methods are developed as a trade-off between accuracy and efficiency. One such method is the INLA methodology. Recently, a fundamental reformulation of the INLA methodology resulted in an even faster and more accurate approximate Bayesian inference framework with wide applicability. In this talk, I will present some case studies where we approach near real-time inference for complex Biostatistics models, such as disease mapping and brain activation mapping models, among others often encountered in the biostatistics domain, using INLA. I will also give a compact overview of the INLA methodology and some insights into the new formulation thereof.
Overview
Presenters
Brief Biography
Janet van Niekerk is a Research Scientist in the Bayesian Computational Statistics and Modeling group. She works on Bayesian methods including prior construction and computational framework development for various applications, with a focus on Biostatistics.
Related People
Related Researchers
Janet van Niekerk
- Research Scientist, Statistics