Curriculum

  /  

Descriptions

MLDS 490: Bayesian Methods for Inference and Decision Making

Description

This course provides a rigorous yet practical introduction to Bayesian methods for statistical inference, probabilistic modeling, knowledge representation, and decision-making under uncertainty. Students will develop a deep understanding of Bayesian reasoning and Bayesian Networks as unified frameworks for learning from data, representing complex domains, discovering causal relationships, and supporting robust decisions in high-dimensional environments.

Topics include Bayesian reasoning using Bayesian networks, analysis of causal relationships, Bayesian inference using conjugate prior families or Markov Chain Monte Carlo, using shrinkage for variable selection and regularization, Bayes factors for model comparison and hypothesis testing.

The course includes extensive hands-on experience with BayesiaLab, a leading software platform for transforming data and expert knowledge into Bayesian Network models that support prediction, diagnosis, simulation, causal reasoning, and decision optimization.

The course is based on a textbook written by the instructor, available through the author's website, together with AI-powered learning tools built around the text that enable students to interactively explore concepts, solve problems, and receive personalized guidance.

By the end of the course, students will be able to build, learn, validate, and apply Bayesian models to perform probabilistic and causal inference and to develop robust, interpretable solutions for complex real-world problems across engineering, healthcare, business, finance, and the social sciences.