Scalable Bayesian Inference for Dynamic Generalized Linear Models
Posterior-Predictive Metrics for Sca-MCMC Fit
Compute Scalable Mutation-Rate Vector
Calculate Log-Likelihood for DGLM
Generate Geometric Inverse-Temperature Ladder Constructs a geometric s...
Generate Geometric Temperature Ladder for Parallel Tempering
Normalized Hamming Distance
Scalable Mutation-Rate Strategies for Sca-MCMC
Print method for SDGLM objects
Print method for summary.SDGLM
Generate Random Samples from the Inverse Wishart Distribution
Scalable MCMC for Dynamic GLMs
Alternative Sca-MCMC Implementation for Variable Selection
SDGLM: Scalable Bayesian Inference for Dynamic Generalized Linear Mode...
Simulate Gamma Dynamic GLM
Simulate Pareto-type Dynamic GLM
Simulate Poisson-Binomial Dynamic GLM
Summary method for SDGLM objects
Implements scalable Markov chain Monte Carlo (Sca-MCMC) algorithms for Bayesian inference in dynamic generalized linear models (DGLMs). The package supports Pareto-type and Gamma-type DGLMs, which are suitable for modeling heavy-tailed phenomena such as wealth allocation and financial returns. It provides simulation tools for synthetic DGLM data, adaptive mutation-rate strategies (ScaI, ScaII, ScaIII), geometric temperature ladders for parallel tempering, and posterior predictive evaluation metrics (e.g., R2, RMSE). The methodology is based on the scalable MCMC framework described in Guo et al. (2025).