Merge branch 'master' of https://github.com/nk027/bvar

NEWS.md

@@ -1,28 +1,40 @@
 # v1.0.0, CRAN Update 5 / JSS Revision 2
 
-add FRED-MD (Update datasets, compress em)
-cond fcast
-bugfixes (irf, vectorised scale_hess)
-improve docs
-add options to plots (area, t_back, x-labels, transparence)
-add vignette (compressed to ebook quality, figures pdf/png)
-add wrapper for parallelisation
-add helper function for transformation
-optimise speed (mvnorm, matrix operations, factor 2 to 10)
-add unit tests
-handle coda interface nicer (zzz.R, fix potential Windows bug)
-remove deprecated functions
-update citation
-add mean argument to methods
-facilitate minnesota prior mean changes
-add replacement functions for irf() and predict()
-move irf and predict out of main loop
-speed up fevd calculation
-add robustness checks
-add helper functions
-improve vars & co to work with regex
-auto psi now tries increasing integration once
-add verbosity to errors
+- Add fancy **vignette** with background and demonstrations
+- Add new features
+  - New **FRED-MD** database and updated FRED-QD (`data("fred_md")`)
+  - Transformation helper functions (`fred_transform()`, `fred_code()`)
+  - Add **Conditional forecasting** (see `?bv_fcast()`)
+  - New replacement functions for `irf()` and `predict()` (`irf(x) <- irf(x)`)
+  - Provide wrapper for parallelised execution (`par_bvar()`)
+- Improve existing features
+  - Enhance IRF and forecast plotting
+    - `area` argument adds polygons for credible intervals
+    - `t_back` allows adding realised values before forecasts
+    - `col` and `fill` arguments allow changing colours
+    - transparence is applied to sequential lines / polygons
+    - Improved x-axis labelling
+  - Regex may be used for `vars`, `vars_response`, and `vars_impulse`
+  - New `type` for `coef()`, `fitted()`, etc, to retrieve means / quantiles
+  - Add constructors for the prior mean `b` argument in `bv_mn()`
+  - Auto `psi` now allows for one order of integration
+- **Enhance speed** considerably (~2-10 times faster)
+  - Move IRF and forecasts out of MCMC
+  - Capitalise upon matrix properties
+  - Cached and customised multivariate normal drawing
+  - Optimised FEVD computation
+- Fix bugs
+  - IRF calculation is now ordered properly (please recalculate)
+  - *coda* methods are now proper methods (potential issue on Windows)
+  - Vectorised `scale_hess` now works properly
+- Remove deprecated functions and arguments
+- Work on documentation and examples
+- Update citation information
+- Improve upon internal structure
+  - Unit tests with *tinytest* for development (skipped on CRAN)
+  - Outsource additional steps to dedicated functions
+  - More robustness checks and add verbosity to errors
+- Tested extensively on R 4.0.0 and R 3.6.3.
 
 
 # v0.2.2, CRAN Update 4 / Impulse Response Hotfix

R/58_fcast_plot.R

@@ -4,8 +4,6 @@
 #' Forecasts of all or a subset of the available variables can be plotted.
 #'
 #' @param x A code{bvar_fcast} object, obtained from \code{\link{predict.bvar}}.
-#' @param conf_bands Deprecated. Use \code{\link{predict.bvar}}. Numeric vector
-#' of desired confidence bands.
 #' @param vars Optional numeric or character vector. Used to subset the plot to
 #' certain variables by position or name (must be available). Defaults to
 #' \code{NULL}, i.e. all variables.
@@ -69,7 +67,6 @@
 #' }
 plot.bvar_fcast <- function(
   x,
-  conf_bands, # deprecated, see `predict.bvar()`
   vars = NULL,
   col = "#737373",
   t_back = 1,
@@ -81,7 +78,7 @@ plot.bvar_fcast <- function(
   ...) {
 
   if(!inherits(x, "bvar_fcast")) {stop("Please provide a `bvar_fcast` object.")}
-  plot_fcast(x = x, conf_bands = conf_bands, vars = vars,
+  plot_fcast(x = x, vars = vars,
     variables = variables, orientation = orientation, mar = mar,
     t_back = t_back, area = area, col = col, fill = fill, ...)
 }
@@ -90,7 +87,6 @@ plot.bvar_fcast <- function(
 #' @noRd
 plot_fcast <- function(
   x,
-  conf_bands, # deprecated, see `predict.bvar()`
   vars = NULL,
   variables = NULL,
   orientation = c("vertical", "horizontal"),
@@ -108,11 +104,6 @@ plot_fcast <- function(
   }
   if(inherits(x, "bvar")) {x <- predict(x)}
 
-  if(!missing(conf_bands)) {
-    message("Parameter conf_bands is deprecated. Please use `predict()`.")
-    x <- predict(x, conf_bands = conf_bands)
-  }
-
   orientation <- match.arg(orientation)
 
   # Prepare data ---

R/68_irf_plot.R

@@ -5,8 +5,6 @@
 #' plotted.
 #'
 #' @param x A \code{bvar_irf} object, obtained from \code{\link{irf.bvar}}.
-#' @param conf_bands Deprecated. Use \code{\link{irf.bvar}}. Numeric vector
-#' of desired confidence bands.
 #' @param vars_impulse,vars_response Optional numeric or character vector. Used
 #' to subset the plot's impulses / responses to certain variables by position
 #' or name (must be available). Defaults to \code{NULL}, i.e. all variables.
@@ -64,7 +62,6 @@
 #' }
 plot.bvar_irf <- function(
   x,
-  conf_bands, # deprecated, see `irf.bvar()`
   vars_response = NULL,
   vars_impulse = NULL,
   col = "#737373",
@@ -75,7 +72,7 @@ plot.bvar_irf <- function(
   ...) {
 
   if(!inherits(x, "bvar_irf")) {stop("Please provide a `bvar_irf` object.")}
-  plot_irf(x = x, conf_bands = conf_bands,
+  plot_irf(x = x,
     vars_response = vars_response, vars_impulse = vars_impulse,
     variables = variables, mar = mar, area = area, col = col, fill = fill, ...)
 }
@@ -84,7 +81,6 @@ plot.bvar_irf <- function(
 #' @noRd
 plot_irf <- function(
   x,
-  conf_bands, # deprecated, see `irf.bvar()`
   vars_response = NULL,
   vars_impulse = NULL,
   variables = NULL,
@@ -101,11 +97,6 @@ plot_irf <- function(
   }
   if(inherits(x, "bvar")) {x <- irf(x)}
 
-  if(!missing(conf_bands)) {
-    message("Parameter conf_bands is deprecated. Please use `irf()`.")
-    x <- irf(x, conf_bands = conf_bands)
-  }
-
   # Prepare data ---
 
   has_quants <- length(dim(x[["quants"]])) == 4L