dsp: An R Package for Bayesian Time Series Modeling with the Dynamic Shrinkage Process
Abstract
dsp provides a unified state-space framework for modeling nonstationary time series using the Dynamic Shrinkage Process (DSP), a class of global-local shrinkage priors. DSP adaptively regularizes time-varying parameters to shrink noise while preserving abrupt changes and other local structure. The package supports Gaussian and count-data smoothing, changepoint detection, time-varying parameter regression, and volatility estimation, with posterior predictive draws available for missing-value imputation and forecasting. Nonstationary time series often contain periods of relative stability punctuated by abrupt changes in mean or variance. dsp offers a common interface for model specification, posterior computation, summarization, and visualization across models that were previously developed separately in the literature. Core functionality is implemented in C++ via Rcpp for computational efficiency. We describe the underlying models, package design, and the main functions for fitting, summarizing, and plotting posterior results.
Related
- Weak Factor Recovery in Sparse Bayesian Factor Models under Exchangeable Shrinkage Process Prior
- Modeling Dynamic Correlation Matrices with Shrinkage Priors
- Structured Bayesian Modeling of Dynamic Receptive1 Fields in Salamander Retinal Ganglion Cells
- Pathwise Shrinkage and Probabilistic Calibration of the Multiplicative Half-Cauchy Process
- Semiparametric Bayesian structure learning of nonparanormal directed acyclic graphs with local--global shrinkage
BibTeX
@misc{cho2026package,
title = {{dsp: An R Package for Bayesian Time Series Modeling with the Dynamic Shrinkage Process}},
author = {Jason B. Cho and Daniel R. Kowal and Haoxuan Wu and David S. Matteson and Toryn L. J. Schafer},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.07096},
eprint = {2611.07096},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.07096},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - dsp: An R Package for Bayesian Time Series Modeling with the Dynamic Shrinkage Process
AU - Jason B. Cho
AU - Daniel R. Kowal
AU - Haoxuan Wu
AU - David S. Matteson
AU - Toryn L. J. Schafer
PY - 2026
DA - 2026/11/04
PB - arXiv
DO - 10.48550/arxiv.2611.07096
UR - https://arxiv.org/abs/2611.07096
AB - dsp provides a unified state-space framework for modeling nonstationary time series using the Dynamic Shrinkage Process (DSP), a class of global-local shrinkage priors. DSP adaptively regularizes time-varying parameters to shrink noise while preserving abrupt changes and other local structure. The package supports Gaussian and count-data smoothing, changepoint detection, time-varying parameter regression, and volatility estimation, with posterior predictive draws available for missing-value imputation and forecasting. Nonstationary time series often contain periods of relative stability punctuated by abrupt changes in mean or variance. dsp offers a common interface for model specification, posterior computation, summarization, and visualization across models that were previously developed separately in the literature. Core functionality is implemented in C++ via Rcpp for computational efficiency. We describe the underlying models, package design, and the main functions for fitting, summarizing, and plotting posterior results.
ER -