pymc - Build and validate Bayesian models with PyMC
Build, fit, validate, compare, and diagnose Bayesian models with PyMC using MCMC, variational inference, predictive checks, LOO, and WAIC.
Tags
Updated: 2026-09-30Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Build Bayesian models
- Fit models with MCMC
- Run variational inference
- Check prior predictions
- Check posterior predictions
- Diagnose sampling issues
- Compare models with LOO
- Compare models with WAIC
- Generate posterior predictions
- Average model predictions
Inputs
- Predictor data
- Outcome data
- Model specifications
- Group structure
- New predictor values
- Random seeds
Outputs
- Fitted inference data
- Diagnostic results
- Posterior summaries
- Predictive distributions
- Model comparison results
- Model averaging weights
- Prediction intervals
- Sampling state changes
Requirements
- Python environment
- PyMC 5.x or later
- ArviZ
- NumPy
- Required diagnostic scripts
- Required comparison scripts