Builds, fits, and validates robust Bayesian models using PyMC's modern probabilistic programming interface.
This skill empowers Claude to perform advanced probabilistic programming tasks using PyMC 5.x+. It provides a comprehensive framework for the entire Bayesian workflow—from data preparation and prior predictive checks to MCMC sampling with NUTS and complex model diagnostics. Whether you're building hierarchical models, time-series analyses, or conducting model comparisons using LOO/WAIC, this skill ensures best practices like non-centered parameterization and proper uncertainty quantification are followed for reliable scientific inference.
主要功能
01Advanced MCMC sampling using the NUTS algorithm and variational inference (ADVI)
022,188 GitHub stars
03Hierarchical and multilevel modeling with automated non-centered parameterization
04Standardized workflows for prior and posterior predictive checks (PPC)
05Comprehensive model diagnostics including R-hat, ESS, and divergence checks
06Model comparison and averaging using LOO and WAIC information criteria
使用场景
01Performing robust time-series forecasting and autoregressive modeling
02Analyzing grouped or nested data using hierarchical Bayesian frameworks
03Quantifying uncertainty in complex linear and logistic regression models
