How does it help with model validation?

The skill follows a strict workflow including prior predictive checks, MCMC convergence diagnostics (R-hat, ESS), and posterior predictive checks to ensure model reliability.

Which version of PyMC does this skill support?

The skill is optimized for PyMC version 5.x and above, utilizing the modern API and ArviZ integration.

Does it support model comparison?

Yes, it provides tools for comparing multiple models using information criteria like Leave-One-Out (LOO) cross-validation and WAIC.

Can this skill handle hierarchical or multi-level data?

Yes, it includes specific patterns for hierarchical models, emphasizing non-centered parameterization to improve sampling efficiency and avoid divergences.

What is the PyMC skill for Claude Code?

This skill provides Claude with domain-specific knowledge to help you build, fit, and validate Bayesian statistical models using the PyMC Python library and ArviZ.

PyMC Bayesian Modeling

Name: PyMC Bayesian Modeling
Author: henriquescastilho

byhenriquescastilho

•

数据科学与机器学习

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

This skill empowers Claude to design, implement, and analyze Bayesian models using the PyMC library. it provides a standardized workflow for probabilistic programming, covering everything from prior selection and data standardization to MCMC sampling and posterior predictive checks. Ideal for handling uncertainty, hierarchical data structures, and complex statistical inference, it includes built-in best practices for model diagnostics and comparison using information criteria like LOO and WAIC.

主要功能

011 GitHub stars

02Model comparison and selection using LOO and WAIC information criteria

03Standardized Bayesian modeling workflow with PyMC 5.x+

04Advanced MCMC sampling (NUTS) and Variational Inference (ADVI) implementation

05Comprehensive model diagnostics including R-hat, ESS, and divergence checks

06Hierarchical and multi-level model templates with non-centered parameterization

使用场景

01Analyzing complex hierarchical or multi-level datasets with group-level effects

02Performing time-series forecasting and autoregressive Bayesian analysis

03Quantifying uncertainty in linear, logistic, and Poisson regression models

主要功能

011 GitHub stars

02Model comparison and selection using LOO and WAIC information criteria

03Standardized Bayesian modeling workflow with PyMC 5.x+

04Advanced MCMC sampling (NUTS) and Variational Inference (ADVI) implementation

05Comprehensive model diagnostics including R-hat, ESS, and divergence checks

06Hierarchical and multi-level model templates with non-centered parameterization

使用场景

01Analyzing complex hierarchical or multi-level datasets with group-level effects

02Performing time-series forecasting and autoregressive Bayesian analysis

03Quantifying uncertainty in linear, logistic, and Poisson regression models