Merge pull request #4 from bsvars/2-gibbs-sampler-for-the-basic-model

2 gibbs sampler for the basic model

DESCRIPTION

@@ -4,17 +4,22 @@ Title: Forecasting with Bayesian Hierarchical Panel Vector Autoregressions
 Description: Forecasting a multi-country time series panel data using Bayesian Vector Autoregressions with a three-level country-global hierarchical prior structure.
 Version: 0.0.1.9000
 Date: 2024-04-27
-Author: c(person(given="Tomasz", family="Woźniak", email="wozniak.tom@pm.me", role = c("aut", "cre"), comment = c(ORCID = "0000-0003-2212-2378")), person(given="Miguel", family="Sanchez-Martinez", role = "ctb"), person(given="International Labour Organization", role = "cph"))
+Author: c(person(given="Tomasz", family="Woźniak", email="wozniak.tom@pm.me", role =
+         c("aut", "cre"), comment = c(ORCID = "0000-0003-2212-2378")),
+         person(given="Miguel", family="Sanchez-Martinez", role = "ctb"),
+         person(given="International Labour Organization", role = "cph"))
 Maintainer: Tomasz Woźniak <wozniak.tom@pm.me>
 License: GPL (>= 3)
+Depends: 
+    R (>= 2.10),
+    bsvars
 Imports: 
     R6,
     Rcpp (>= 1.0.12),
     RcppProgress,
     RcppTN
 LinkingTo: bsvars, Rcpp, RcppArmadillo, RcppProgress, RcppTN
+Suggests: tinytest
 LazyData: true
 Encoding: UTF-8
 RoxygenNote: 7.3.1
-Depends: 
-    R (>= 2.10)

NAMESPACE

@@ -1,5 +1,16 @@
 # Generated by roxygen2: do not edit by hand
 
+S3method(estimate,BVARPANEL)
+S3method(estimate,PosteriorBVARPANEL)
+export(specify_bvarPANEL)
+export(specify_panel_data_matrices)
+export(specify_posterior_bvarPANEL)
+export(specify_prior_bvarPANEL)
+export(specify_starting_values_bvarPANEL)
+import(RcppProgress)
 importFrom(R6,R6Class)
 importFrom(Rcpp,sourceCpp)
+importFrom(RcppTN,dtn)
+importFrom(RcppTN,rtn)
+importFrom(bsvars,estimate)
 useDynLib(bvarPANELs, .registration = TRUE)

NEWS.md

@@ -0,0 +1,5 @@
+# bvarPANELs 0.0.1.9000
+
+1. This package is alive! [#1](https://github.com/bsvars/bvarPANELs/issues/1)
+2. The package includes the Gibbs sampler for a Bayesian Hierarchical Panel Vector Autoregression model. [#2](https://github.com/bsvars/bvarPANELs/issues/2)
+3. The package includes labour market dynamic panel data. [#3](https://github.com/bsvars/bvarPANELs/issues/3)

R/RcppExports.R

@@ -1,7 +1,43 @@
 # Generated by using Rcpp::compileAttributes() -> do not edit by hand
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
+bvarPANEL <- function(S, Y, X, prior, starting_values, thin = 100L, show_progress = TRUE) {
+    .Call(`_bvarPANELs_bvarPANEL`, S, Y, X, prior, starting_values, thin, show_progress)
+}
+
 rmniw1 <- function(A, V, S, nu) {
     .Call(`_bvarPANELs_rmniw1`, A, V, S, nu)
 }
 
+sample_m <- function(aux_A, aux_V, aux_s, aux_w, prior) {
+    .Call(`_bvarPANELs_sample_m`, aux_A, aux_V, aux_s, aux_w, prior)
+}
+
+sample_w <- function(aux_V, prior) {
+    .Call(`_bvarPANELs_sample_w`, aux_V, prior)
+}
+
+sample_s <- function(aux_A, aux_V, aux_Sigma, aux_m, prior) {
+    .Call(`_bvarPANELs_sample_s`, aux_A, aux_V, aux_Sigma, aux_m, prior)
+}
+
+log_kernel_nu <- function(aux_nu, aux_Sigma_c, aux_Sigma, prior_lambda, C, N, K) {
+    .Call(`_bvarPANELs_log_kernel_nu`, aux_nu, aux_Sigma_c, aux_Sigma, prior_lambda, C, N, K)
+}
+
+sample_nu <- function(aux_nu, aux_Sigma_c, aux_Sigma, prior) {
+    .Call(`_bvarPANELs_sample_nu`, aux_nu, aux_Sigma_c, aux_Sigma, prior)
+}
+
+sample_Sigma <- function(aux_Sigma_c_inv, aux_s, aux_nu, prior) {
+    .Call(`_bvarPANELs_sample_Sigma`, aux_Sigma_c_inv, aux_s, aux_nu, prior)
+}
+
+sample_AV <- function(aux_A_c, aux_Sigma_c_inv, aux_s, aux_m, aux_w, prior) {
+    .Call(`_bvarPANELs_sample_AV`, aux_A_c, aux_Sigma_c_inv, aux_s, aux_m, aux_w, prior)
+}
+
+sample_A_c_Sigma_c <- function(Y_c, X_c, aux_A, aux_V, aux_Sigma, aux_nu) {
+    .Call(`_bvarPANELs_sample_A_c_Sigma_c`, Y_c, X_c, aux_A, aux_V, aux_Sigma, aux_nu)
+}
+

R/bvarPANELs-package.R

@@ -30,9 +30,18 @@
 #' @aliases bvarPANELs-package bvarPANELs
 #' @docType package
 #' @useDynLib bvarPANELs, .registration = TRUE
+#' @importFrom bsvars estimate
 #' @importFrom Rcpp sourceCpp
 #' @importFrom R6 R6Class
+#' @importFrom RcppTN rtn dtn
+#' @import RcppProgress
 #' @note This package is currently in active development.
 #' @author Tomasz Woźniak \email{wozniak.tom@pm.me}
 #' @keywords package models ts
+#' #' @examples
+#' data(ilo_cubic_panel)                                   # load the data
+#' set.seed(123)
+#' specification = specify_bvarPANEL$new(ilo_cubic_panel)  # specify the model
+#' burn_in       = estimate(specification, 10)             # run the burn-in
+#' posterior     = estimate(burn_in, 10)                   # estimate the model
 "_PACKAGE"

R/estimate.bvarPANEL.R

@@ -0,0 +1,94 @@
+
+#' @title Bayesian estimation of a Bayesian Hierarchical Panel Vector 
+#' Autoregression using Gibbs sampler
+#'
+#' @description Estimates the Bayesian Hierarchical Panel VAR using the Gibbs 
+#' sampler proposed by Woźniak (2024).
+#' 
+#' @details 
+#' The homoskedastic SVAR model is given by the reduced form equation:
+#' \deqn{Y_c = A_cX_c + E_c}
+#' where \eqn{Y_c} is an \code{T_c x N} matrix of dependent variables for 
+#' country \code{c}, \eqn{X_c} is a \code{T_c x K} matrix of explanatory 
+#' variables, \eqn{E_c} is an \code{T_c x N} matrix of error terms, and 
+#' \eqn{A_c} is an \code{NxK} matrix of country-specific autoregressive slope 
+#' coefficients and parameters on deterministic terms in \eqn{X_c}.
+#' 
+#' @param specification an object of class BVARPANEL generated using the 
+#' \code{specify_bvarPANEL$new()} function.
+#' @param S a positive integer, the number of posterior draws to be generated
+#' @param thin a positive integer, specifying the frequency of MCMC output thinning
+#' @param show_progress a logical value, if \code{TRUE} the estimation progress 
+#' bar is visible
+#' 
+#' @return An object of class PosteriorBVARPANEL containing the Bayesian 
+#' estimation output and containing two elements:
+#' 
+#'  \code{posterior} a list with a collection of \code{S} draws from the 
+#'  posterior distribution generated via Gibbs sampler.
+#' \code{last_draw} an object of class BVARPANEL with the last draw of the 
+#' current MCMC run as the starting value to be passed to the continuation of 
+#' the MCMC estimation using the \code{estimate()} method. 
+#'
+#' @seealso \code{\link{specify_bvarPANEL}}, \code{\link{specify_posterior_bvarPANEL}}
+#'
+#' @author Tomasz Woźniak \email{wozniak.tom@pm.me}
+#' 
+#' @method estimate BVARPANEL
+#' 
+#' @examples
+#' data(ilo_cubic_panel)                                   # load the data
+#' set.seed(123)
+#' specification = specify_bvarPANEL$new(ilo_cubic_panel)  # specify the model
+#' burn_in       = estimate(specification, 10)             # run the burn-in
+#' posterior     = estimate(burn_in, 10)                   # estimate the model
+#' 
+#' @export
+estimate.BVARPANEL <- function(
+    specification, 
+    S, 
+    thin = 10L, 
+    show_progress = TRUE
+) {
+
+  # get the inputs to estimation
+  prior               = specification$prior$get_prior()
+  starting_values     = specification$starting_values$get_starting_values()
+  data_matrices       = specification$data_matrices$get_data_matrices()
+
+  # estimation
+  qqq                 = .Call(`_bvarPANELs_bvarPANEL`, S, data_matrices$Y, data_matrices$X, prior, starting_values, thin, show_progress)
+
+  specification$starting_values$set_starting_values(qqq$last_draw)
+  output              = specify_posterior_bvarPANEL$new(specification, qqq$posterior)
+
+  return(output)
+} # END estimate.BVARPANEL
+
+
+
+#' @inherit estimate.BVARPANEL
+#' 
+#' @method estimate PosteriorBVARPANEL
+#' 
+#' @param specification an object of class PosteriorBVARPANEL generated using 
+#' the \code{estimate.BVARPANEL()} function. This setup facilitates the 
+#' continuation of the MCMC sampling starting from the last draw of the previous 
+#' run.
+#' 
+#' @export
+estimate.PosteriorBVARPANEL <- function(specification, S, thin = 10, show_progress = TRUE) {
+
+  # get the inputs to estimation
+  prior               = specification$last_draw$prior$get_prior()
+  starting_values     = specification$last_draw$starting_values$get_starting_values()
+  data_matrices       = specification$last_draw$data_matrices$get_data_matrices()
+
+  # estimation
+  qqq                 = .Call(`_bvarPANELs_bvarPANEL`, S, data_matrices$Y, data_matrices$X, prior, starting_values, thin, show_progress)
+
+  specification$last_draw$starting_values$set_starting_values(qqq$last_draw)
+  output              = specify_posterior_bvarPANEL$new(specification$last_draw, qqq$posterior)
+
+  return(output)
+} # END estimate.PosteriorBSVAR