Can I perform model comparison with this skill?

Yes, the skill includes templates for comparing multiple models using Leave-One-Out (LOO) cross-validation and WAIC, including Pareto-k reliability checks.

Which version of PyMC is supported by this skill?

This skill is specifically designed for PyMC version 5.x and later, utilizing the most recent API patterns and the ArviZ library for posterior analysis.

How does the skill handle MCMC sampling divergences?

The skill provides specific troubleshooting steps for divergences, including increasing target_accept values and implementing non-centered parameterization for hierarchical models.

Does it support Variational Inference (VI)?

Yes, it provides patterns for using Automatic Differentiation Variational Inference (ADVI) for quick exploration and initialization of MCMC chains.

PyMC Bayesian Modeling

Name: PyMC Bayesian Modeling
Author: BbgnsurfTech

byBbgnsurfTech

•

Ciencia de Datos y ML

Builds, fits, and validates robust Bayesian statistical models using the PyMC probabilistic programming framework.

Acerca de

This skill provides a comprehensive workflow for performing Bayesian inference within Claude Code using the PyMC library. It guides users through the entire probabilistic programming lifecycle, including data preparation, weakly informative prior selection, hierarchical model construction, and high-performance MCMC sampling via NUTS. The skill emphasizes best practices like non-centered parameterization for multilevel models, rigorous diagnostic checks using R-hat and ESS, and model comparison through information criteria like LOO and WAIC, ensuring statistically sound results for complex data analysis tasks.

Características Principales

Support for hierarchical and non-centered parameterization
Model comparison and selection using LOO/WAIC metrics
Standardized Bayesian workflow from priors to posterior checks
Integrated diagnostic scripts for convergence and sampling issues
3 GitHub stars
Automated MCMC (NUTS) and Variational Inference (ADVI) setups

Casos de Uso

Analyzing time-series data with autoregressive Bayesian processes
Modeling multi-level grouped data with hierarchical structures
Estimating uncertainty in complex linear and logistic regressions

Acerca de

Características Principales

Support for hierarchical and non-centered parameterization
Model comparison and selection using LOO/WAIC metrics
Standardized Bayesian workflow from priors to posterior checks
Integrated diagnostic scripts for convergence and sampling issues
3 GitHub stars
Automated MCMC (NUTS) and Variational Inference (ADVI) setups

Casos de Uso

Analyzing time-series data with autoregressive Bayesian processes
Modeling multi-level grouped data with hierarchical structures
Estimating uncertainty in complex linear and logistic regressions