CMPINF 2221 APPLIED BAYESIAN ANALYSIS

In this course students will learn the foundational assumptions, concepts, and popular tools for applying Bayesian techniques to solve challenging data related problems. This course is hands-on and demonstrates the key concepts and skills through numerous programming examples in lectures, homework assignments, and projects. The course begins by reviewing probability distributions with a special emphasis on how distributions communicate uncertainty. The Bayesian "mindset" is introduced by showing how probability distributions allow subjective information to be used in modeling tasks via Prior distributions. Students will be shown the connection between Bayesian Priors and Non-Bayesian regularization/penalization methods which they have been introduced to in previous courses. From there, students are taught how to properly train, validate, and communicate the Bayesian modeling results for linear, generalized linear models, and multi-level (hierarchical) models using popular open-source libraries. Special emphasis is made to diagnose the Bayesian inference procedure to ensure the models are adequate and trustworthy.

Academic Career: Graduate
Course Component: Lecture
Grade Component: Grad LG/SNC Basis
Minimum Credits: 3
Maximum Credits: 3

Current Sections