One-click install
npx skills add https://github.com/HaykTarkhanyan/dst_research --skill pymc-hayktarkhanyan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pymc
Source: https://github.com/HaykTarkhanyan/dst_research/tree/main/.claude/skills/pymc
Command: npx skills add https://github.com/HaykTarkhanyan/dst_research --skill pymc-hayktarkhanyan

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) and references (resource) components.

What problem does it solve?

Bayesian modeling and probabilistic programming requires careful workflow to build, fit, and validate models with quantified uncertainty. This skill provides a practical, end-to-end blueprint for employing PyMC to design hierarchical models, perform MCMC or variational inference, and evaluate models with robust posterior checks.

Core Features & Use Cases

  • Build and fit Bayesian models (linear, logistic, hierarchical, time-series) with PyMC 5.x+.
  • Perform diagnostic checks (R-hat, ESS, divergences) and posterior predictive checks with ArviZ.
  • Compare models using LOO/WAIC and apply model-averaging strategies across disciplines.

Quick Start

Define your data, build a PyMC model, run MCMC or VI, and review diagnostics to interpret the posterior results.

Frequently Asked Questions about pymc

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build and validate Bayesian hierarchical models with PyMC?

Build Bayesian models with PyMC by defining data, priors, and likelihood, then validate using MCMC sampling and ArviZ diagnostics like R-hat, ESS, and divergences for accurate posterior results.

What is the best way to compare Bayesian models using LOO and WAIC?

Compare Bayesian models using LOO and WAIC by computing information criteria with ArviZ, enabling model selection and model-averaging strategies to evaluate predictive accuracy across disciplines.

How does MCMC sampling with NUTS work for probabilistic programming?

MCMC sampling with NUTS works by exploring posterior distributions in probabilistic programming, generating samples that quantify uncertainty for Bayesian models including regression, time-series, and hierarchical structures.

Can I use variational inference instead of MCMC for Bayesian regression?

Use variational inference as an alternative to MCMC for Bayesian regression when faster approximation is needed, trading exact posterior exploration for computational efficiency in PyMC workflows.

Do I need ArviZ to perform posterior predictive checks?

ArviZ is required to perform posterior predictive checks, providing diagnostic plots and statistical summaries to validate model fit and assess whether simulated data matches observed distributions.

Why are divergences appearing in my MCMC sampling diagnostics?

Divergences in MCMC sampling diagnostics indicate problematic posterior geometry, often caused by inappropriate priors or hierarchical model parameterization, requiring model reparameterization or prior adjustment.