PyMC Bayesian Modeling

소개

This skill equips Claude with specialized knowledge for implementing comprehensive Bayesian workflows in Python using the PyMC library. It provides expert guidance on constructing complex hierarchical models, performing MCMC sampling with NUTS, and conducting rigorous model diagnostics via ArviZ. Whether you are dealing with uncertainty quantification, missing data, or multilevel structures, this skill ensures best practices like non-centered parameterization and prior/posterior predictive checks are followed to achieve robust, interpretable inference.

주요 기능

• Hierarchical and multilevel model construction with non-centered parameterization
• Prior and posterior predictive checks for rigorous model validation
• 0 GitHub stars
• Comprehensive model diagnostics including R-hat, ESS, and divergence checking
• Advanced MCMC sampling (NUTS) and variational inference (ADVI) implementation
• Model comparison and selection using LOO and WAIC criteria

사용 사례

• Estimating uncertainty in complex linear and logistic regression models
• Analyzing grouped or nested data using hierarchical Bayesian structures
• Forecasting and time series analysis using autoregressive processes