Stan Bayesian Modeling
Purpose
Provides Stan probabilistic programming capabilities for Bayesian inference and statistical modeling.
Capabilities
- Stan model specification
- MCMC sampling (NUTS, HMC)
- Variational inference
- Prior predictive checks
- Posterior predictive checks
- Model comparison (LOO-CV, WAIC)
Usage Guidelines
- Model Specification: Write Stan code with clear blocks
- Prior Selection: Choose appropriate, weakly informative priors
- Diagnostics: Check Rhat, ESS, and divergences
- Model Comparison: Use LOO-CV for model selection
Tools/Libraries
- Stan
- CmdStan
- RStan
- PyStan