Can I compare different Bayesian models using this skill?

Yes, it includes standardized patterns for model comparison using information criteria such as LOO (Leave-One-Out cross-validation) and WAIC.

How does this skill help with sampling issues like divergences?

It provides specific guidance on reparameterization techniques, such as non-centered parameterization, and suggests tuning parameters like target_accept to resolve NUTS sampler divergences.

Does it support uncertainty quantification for predictions?

Absolutely. The skill facilitates full posterior predictive checks and provides methods to extract Highest Density Intervals (HDI) for principled uncertainty quantification.

What version of PyMC does this skill support?

This skill is optimized for modern PyMC (version 5.x+) and leverages the latest APIs for model building, sampling, and ArviZ integration.

PyMC Bayesian Modeling

Name: PyMC Bayesian Modeling
Author: Microck

byMicrock

•

데이터 과학 및 ML

Builds and validates sophisticated Bayesian models using PyMC for probabilistic programming and statistical inference.

소개

This skill empowers Claude to architect, implement, and diagnose complex Bayesian models within the PyMC framework. It provides a standardized workflow covering data preparation, prior predictive checks, MCMC sampling using the NUTS sampler, and rigorous model validation through posterior checks and information criteria like LOO and WAIC. It is particularly useful for handling hierarchical data, uncertainty quantification, and implementing non-centered parameterizations to ensure robust sampling and reliable statistical inference.

주요 기능

Patterns for hierarchical modeling and non-centered parameterization
Comprehensive Bayesian workflow from data preparation to posterior analysis
81 GitHub stars
Automated model diagnostics including R-hat, ESS, and divergence checks
Advanced model comparison using LOO and WAIC criteria
Support for MCMC sampling (NUTS) and Variational Inference (ADVI)

사용 사례

Performing time series analysis and autoregressive modeling within a probabilistic framework
Building robust regression models with weakly informative or domain-specific priors
Analyzing multi-level or hierarchical datasets with principled uncertainty quantification

소개

주요 기능

Patterns for hierarchical modeling and non-centered parameterization
Comprehensive Bayesian workflow from data preparation to posterior analysis
81 GitHub stars
Automated model diagnostics including R-hat, ESS, and divergence checks
Advanced model comparison using LOO and WAIC criteria
Support for MCMC sampling (NUTS) and Variational Inference (ADVI)

사용 사례

Performing time series analysis and autoregressive modeling within a probabilistic framework
Building robust regression models with weakly informative or domain-specific priors
Analyzing multi-level or hierarchical datasets with principled uncertainty quantification