Bayesian modeling with PyMC. Build hierarchical models, MCMC (NUTS), variational inference, LOO/WAIC comparison, posterior checks, for probabilistic programming and inference.
The pymc-bayesian-modeling skill provides a comprehensive framework for building and validating Bayesian models using the PyMC library. This skill addresses the challenges associated with probabilistic programming by offering tools to create hierarchical models, perform MCMC sampling using the Hamiltonian Monte Carlo method (NUTS), and conduct variational inference. It simplifies the process of model comparison through techniques like Leave-One-Out (LOO) and Widely Applicable Information Criterion (WAIC), enabling users to carry out rigorous posterior checks and diagnostics effectively.
You can build various types of Bayesian models, including linear and logistic regression, hierarchical models, and time series.
This skill is designed for PyMC version 5.x and above, so compatibility with earlier versions may not be guaranteed.
You need to have Python and the PyMC library, along with its dependencies, installed in your environment to use this skill.
Yes, the skill allows for principled handling of missing data by treating it as parameters.
Yes, the skill includes diagnostic functionalities to check for divergences, convergence, and effective sample size (ESS).
Quick Setup:
.claude/skills/Repository
davila7/claude-code-templates