Astrophysics/Space Physics Seminar – Dr. Robert Mutel and co-speaker Will Golay | Physics

Bayesian Inference: An Illustrated Guide for Astronomers

Dr. Robert Mutel with co-speaker Will Golay; University of Iowa

Bayesian statistical methods have experienced a dramatic growth in astronomical research in the past two decades. Why is this? There are two compelling reasons: The ability to easily include ‘prior’ information, and better characterization of model parameter uncertainties. We will illustrate Bayesian inference with two simple ‘toy’ problems that compare traditional (frequentist) and Bayesian analyses. Next we will briefly describe how model parameter uncertainties (the posterior distribution) are calculated numerically, a very CPU-intensive task with many options. Finally, we will describe the application of Bayeisian inference in two non-trivial research projects: Exoplanet characterization in short-period binaries, and plasma parameters derived from stellar radio sources.