Regression Analysis Using Bayesian Statistics (2016)

Undergraduate: Kevin Anderson

Faculty Advisor: Christian Iliadis
Department: Physics & Astronomy

---

Traditional methods of statistical analysis are often less flexible than more powerful Bayesian methods. In this talk, I briefly discuss the foundations of Bayesian statistics and show how to construct Bayesian regression models for data analysis. I first introduce basic linear regression models and then demonstrate the ease of use and flexibility provided by Bayesian methods in situations where traditional statistical techniques are of little use (combining discrepant datasets, including measurement uncertainty in parameter estimation, etc). The models are implemented in the JAGS model description language and used within the R programming environment.