Bayesian Health Data Science

Catalog description

This course will teach students how to analyze biomedical data from a Bayesian inference perspective with a strong emphasis on using real-world data, including electronic health records, wearables, and imaging data. The course will begin by introducing the machinery of Bayesian statistics through the lens of linear regression, giving enough context for students with no prior experience with Bayesian statistics. A history of computational approaches used in Bayesian statistics will be given before ultimately landing on Stan, a state-of-the-art probabilistic programming language that makes Bayesian inference accessible as a viable data science tool. The course will then branch out from regression and introduce Bayesian versions of machine learning tools, including regularization and classification. The course will then emphasize Bayesian hierarchical models, including Gaussian process models for temporal and spatial data; and clustering. Additional topics may be discussed from the Bayesian perspective, including causal inference, and meta-analysis. While an applied course, the methods will be introduced from a mathematical perspective, allowing students to obtain a fundamental understanding of the introduced models. Students will learn computational skills for implementing Bayesian models using R and Stan. By the end of this course, students will be well-equipped to tackle complex problems in biomedical research using Bayesian inference. Credits: 3.