Introduction to PyMC3 for Bayesian Modeling and Inference

Welcome to Course 3! This module introduces learners to probabilistic programming with PyMC3, covering the basics, composition of distributions, credible and confidence intervals, posterior predictive checks, robust and hierarchical models, and more. Participants will also gain an understanding of probabilistic programming frameworks and plate notation.

This module delves into linear regression, including mean-centering, robustness, hierarchical and polynomial linear regression, as well as logistic regression and decision boundary for classification. It also includes multiple and multiclass logistic regression, and case studies to reinforce the concepts. Additionally, learners will explore PyMC3 further in this module.

Metrics in PyMC3 module covers various aspects such as Metropolis and HMC, mixing, potential scale reduction factor, centered and non-centered parameterization, convergence assessment, forest plots for visualization, Monte Carlo error, divergences, diagnosing issues, debugging, visualization in Bayesian workflow, and tuning. It also introduces improved Rhat for model assessment and explores PyMC3 in depth.

The final module focuses on modeling COVID-19 cases using PyMC3, providing learners with practical experience in applying the knowledge and techniques acquired throughout the course to a real-world scenario. Participants will gain hands-on experience in modeling and analyzing COVID-19 data using PyMC3, further solidifying their understanding and skills.