Changes in version 3.1.1 (2026-01-08)                  

  - Fixed a bug in the method for calculating fitted values for dynamic
    shrinkage models when only one covariate is present
      - Fixed issues with links to functions in other packages in the
        documentation
      - Fixed various issues with links in the vignette

                 Changes in version 3.1.0 (2025-06-02)                  

  - Exported all dynamic steps via inst/include/shrinkTVP.h
      - Fixed bug in forecasting functions for dynamic shrinkage
      - Fixed bug that prevented shaded argument in plot.mcmc.tvp from
        working

                        Changes in version 3.0.2                        

  - Some updates in the documentation, particularly references to
    multivariate extension shrinkTVPVAR
      - Moved to newer roxygen2 version
      - Added on.exit() to better handle user par settings

                 Changes in version 3.0.1 (2024-02-18)                  

  - Added capability to estimate models under the dynamic triple gamma
    prior (as in Knaus and Frühwirth-Schnatter
    \doi{10.48550/arXiv.2312.10487})
      - Fixed a bug in resample_alpha
      - Fixed an issue on Debian

                 Changes in version 2.1.1 (2023-11-22)                  

  - Added a more robust solve method for armadillo matrices
      - Fixed issues with exported Bhattacharya and Rue algorithm
      - Fixed issue with the usenames option for xcolor in the vignette

                 Changes in version 2.0.6 (2022-10-18)                  

  - Fixed a bug in LPDS()
      - Fixed issues on Fedora and Debian
      - Changed compiler flag -DARMA_DONT_PRINT_ERRORS to
        -DARMA_WARN_LEVEL=0

                 Changes in version 2.0.5 (2022-02-03)                  

  - Modified shrinkTVP.ltx to work with new jss.cls

                 Changes in version 2.0.4 (2021-11-26)                  

  - Added CITATION file
      - Modified DESCRIPTION as well as some documentation files

                 Changes in version 2.0.3 (2021-10-21)                  

  - Fixed a bug in the sampling step for c_xi (Thanks to Tony Chernis
    for pointing it out)
      - Fixed a bug in the sampling step for a_tau (Thanks to Wenjie
        Zhao for pointing it out)
      - Added DG_MH_step to exported functions
      - Fixed some typos in the documentation
      - Added a newline to the print() function

                 Changes in version 2.0.2 (2021-05-13)                  

  - Moved to stochvol 3.0
      - Fixed an issue with the exported version sample_TG_TVP function
      - Fixed an issue with the exported Bhattacharya algorithm
      - define STRICT_R_HEADERS, include float.h, adjust some constants

                 Changes in version 2.0.1 (2020-11-09)                  

  - Fixed an issue on Solaris

                  Changes in version 2.0 (2020-11-05)                   

  - Added option to use triple gamma prior
      - Added one-step update function updateTVP for modular use in
        other samplers
      - Added convenience methods (residuals.shrinkTVP,
        fitted.shrinkTVP, etc...)
      - Re-factored C++ code, improving on speed in some places
      - Added forecast_shrinkTVP (and associated plot.forecast_shrinkTVP
        method), which allows for forecasting
      - All C++ functions are now exported for use in other packages
      - simTVP now allows user to adjust parameters governing SV process
      - Re-factored plot.shrinkTVP to use layout() instead of par(mfrow
        = ...)
      - Bug fixes
      - Updated the vignette

                 Changes in version 1.1.1 (2019-10-06)                  

  - Updated the vignette.

                 Changes in version 1.1.0 (2019-10-06)                  

  - Overhauled plotting functions. The default is now shaded areas
    instead of lines and more customization options are available.
      - plot.mcmc.tvp now also adds time series information to the
        x-axis labels, thanks to zoo::index.
      - Changed the way LPDS are calculated, moving them from the
        shrinkTVP function to a dedicated LPDS function.
      - Added the new eval_pred_dens function, which allows the user to
        evaluate the one-step ahead predictive density.
      - Objects returned from the shrinkTVP function are now of class
        shrinkTVP and not shrinkTVP_res.

                 Changes in version 1.0.2 (2019-08-07)                  

  - Added a vignette
      - Bug fixes

                 Changes in version 1.0.1 (2019-07-15)                  

  - plot.mcmc.tvp now adds a horizontal line at 0
      - Added a nicer print method for shrinkTVP_res objects
      - Fixed some typos and errors in the documentation