ProductPartitionModels.jl implements Gaussian PPMx models (for continuous response and covariates) including imputation-free local regression on partially observed covariates. See the article for details. Code for fitting and assessing the models in R is found in VDLocalReg_examples.
Reference
"A Projection Approach to Local Regression with Variable-Dimension Covariates," Heiner, M., Page, G. and Quintana, F. (2023+) arXiv preprint (link to paper)
Bayesian Nonparametric Density Autoregression
Julia
BNP_WMReg_Joint.jl provides implementation of a nonparametric mixture of autoregressive models with lag selection. See the article for details. Code for fitting and assessing the models is found in BNP_WMAR_examples.
Reference
Bayesian Nonparametric Density Autoregression with Lag Selection, Heiner, M. and Kottas, A. (2022), Bayesian Analysis. (link to paper)
Mixture Transition Distribution (MTD)
Julia
MTD.jl provides a Bayesian implementation of the MTD model, along with our extension for model selection. See the article for details. Code for fitting and assessing the models is found in MTD_examples.
Reference
Estimation and Selection for High-Order Markov Chains with Bayesian Mixture Transition Distribution Models, Heiner, M. and Kottas, A. (2022), Journal of Computational and Graphical Statistics. (link to paper)
Sparse Probability Vectors
Julia
SparseProbVec provides functions to sample the sparse Dirichlet mixture and stick-breaking mixture distributions for probability vectors. See the article for details.
R
SparseProbVec is an R package. See the test script and function documentation for usage examples.
Reference
Structured priors for sparse probability vectors with application to model selection in Markov chains, Heiner, M., Kottas, A., and Munch, S. (2019), Statistics and Computing. (link to paper)