pymc

To build Bayesian models, define priors and likelihoods using PyMC APIs, then run NUTS or Variational Inference sampling to obtain posterior estimates. The Skill supports the end-to-end workflow from model specification to MCMC sampling.

The best way to compare Bayesian models is using LOO or WAIC criteria. This Skill supports model evaluation and comparison by providing diagnostics, visualization, and posterior predictive checks to validate and select among alternative models.

Yes, you can use PyMC for hierarchical and time-series analysis. The Skill applies probabilistic programming to multilevel models and time-series data, enabling robust uncertainty quantification and principled Bayesian regression.

To perform prior and posterior predictive checks, use the diagnostics and plotting utilities provided. The Skill supports validating Bayesian models by evaluating prior and posterior predictive distributions to ensure model fit.

Yes, ArviZ and NumPy are required dependencies. The Skill relies on ArviZ for diagnostics and plotting, NumPy for numerical operations, and Pandas and Matplotlib for data handling and visualization within the probabilistic programming workflow.

To specify priors and use coordinates in Bayesian inference, utilize PyMC's data containers and dims APIs. The Skill supports defining model dimensions, priors, and likelihoods to structure probabilistic models effectively before sampling.