birdie 0.6.1

• Switch to an inverse-gamma prior for the random effects scale parameter in the mixed model.
• Add Jacobian adjustment to mixed model M step to improve accuracy and reduce prior sensitivity.

birdie 0.5.0

• Gibbs sampler for Categorical-Dirichlet and Normal linear models
• simulate.bisg() to perform multiple imputation from BISG probabilities

birdie 0.4.0

• Support for continuous outcomes
• Support for observation/likelihood weights
• New method for specifying outcome model type
• New default empirical Bayes prior for Categorical-Dirichlet model

birdie 0.3.0

• New interfaces and implementations for BISG
• New computational backend and interface for BIRDiE models
• Updated documentation and examples
• Unit tests for accuracy and usability

birdie 0.1.0

• Basic BISG functionality and initial model support