- Update description

.gitignore

@@ -3,5 +3,6 @@
 .RData
 .Ruserdata
 .httr-oauth
+.cran-comments.md
 *.o
 *.so

DESCRIPTION

@@ -10,7 +10,7 @@ Authors@R: c(
 Maintainer: Philipp Adämmer <adaemmer@hsu-hh.de>
 BugReports: https://github.com/adaemmerp/lpirfs/issues
 Description: Contains functions to estimate linear and nonlinear impulse 
-             responses based on local projections by Jordà (2005) <doi:10.1257/0002828053828518>
+             responses based on local projections by Jordà (2005) <doi:10.1257/0002828053828518>.
              Nonlinear impulse responses are estimated for two regimes based on 
              a transition function used in Auerbach and Gorodnichenko (2012) <doi:10.1257/pol.4.2.1>.
 License: GPL (>= 2)
@@ -23,12 +23,8 @@ Imports:
     dplyr(>= 0.7.4),
     foreach(>= 1.4.3),
     ggplot2(>= 2.2.1),
-    ggthemes(>= 3.4.0),
-    gtable(>= 0.2.0),
     mFilter(>= 0.1-3),
     Rcpp(>= 0.12.17),
-    rlang(>= 0.2.0), 
-    rmarkdown(>= 1.9), 
     vars(>= 0.7.4)
 RoxygenNote: 6.0.1
 LinkingTo: Rcpp, RcppArmadillo

R/RcppExports.R

@@ -2,7 +2,7 @@
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
 #' @name find_lag_c
-#' @title Determine optimal lag length
+#' @title Compute optimal lag length
 #' @description Determine optimal lag length based on 'AICc', 'AIC', or 'BIC' criterion.
 #'
 #' @param y List with left (endogenous) variables

R/create_lags.R

@@ -1,5 +1,5 @@
 #' @name create_lags
-#' @title Create a data frame with lagged exogenous variables
+#' @title Compute data frame with lagged exogenous variables
 #' @description Create a data frame with lagged exogenous variables
 #' @param data A data.frame()
 #' @param lags Integer for the number of lags

R/create_lin_data.R

@@ -1,5 +1,5 @@
 #' @name create_lin_data
-#' @title Function to create data for linear model
+#' @title Compute data for linear model
 #' @description Function to create data for linear model.
 #' @param specs A \link{list}(). Inputs are outlined in \link{lp_lin}().
 #' @param data_set_df A \link{data.frame}() with all endogenous variables.

R/create_nl_data.R

@@ -1,5 +1,5 @@
 #' @name create_nl_data
-#' @title Function to create data for nonlinear model
+#' @title Compute data for nonlinear model
 #' @description Function to create data for non-linear model
 #' @param specs A \link{list}(). Inputs are outlined in \link{lp_nl}
 #' @param data_set_df A data frame with all endogenous variables

R/data-documentation.R

@@ -1,4 +1,4 @@
-#' A data frame to estimate the effects of interest rate rules for monetary policy
+#' A data frame containing data to estimate the effects of interest rate rules for monetary policy
 #'
 #' @format A data frame with 193 quarterly observations (rows) and 3 variables (columns):
 #' \describe{
@@ -11,7 +11,7 @@
 "interest_rules_var_data"
 
 
-#'  A data frame to estimate a standard monetary VAR
+#'  A data frame containing data to estimate a standard monetary VAR
 #'
 #'
 #' @format A data frame with 494 monthly observations (rows) and 6 variables (clumns):

R/lp_lin.R

@@ -6,10 +6,10 @@
 #'                    is used for the Cholesky decomposition.
 #' @param lags_criterion NaN or character. NaN means that the number of lags
 #'         will be given at \emph{lags_lin}. The character refers to the corresponding lag length criterion ('AICc', 'AIC' or 'BIC').
-#' @param lags_lin NaN or integer. Number of lags for (linear) VAR (if \emph{lags_criterion} = NaN). NaN if lags_criterion is given.
-#' @param max_lags NaN or integer. Maximum number of lags (if \emph{lags_criterion} = 'AICc', 'AIC', 'BIC'). NaN otherwise.
+#' @param lags_lin NaN or integer. NaN if lag length criterion is used. Integer for number of lags for linear VAR.
+#' @param max_lags NaN or integer. Maximum number of lags if \emph{lags_criterion} is character with lag length criterion. NaN otherwise.
 #' @param trend Integer. No trend =  0 , include trend = 1, include trend and quadratic trend = 2.
-#' @param shock_type Integer. Standard deviation shock = 0, Unit shock = 1.
+#' @param shock_type Integer. Standard deviation shock = 0, unit shock = 1.
 #' @param confint Double. Width of confidence bands. 68\% = 1, 90\% = 1.65, 95\% = 1.96.
 #' @param hor Integer. Number of horizons for impulse responses.
 #'
@@ -20,7 +20,7 @@
 #'                    The last dimension denotes the shock variable. The row in each matrix
 #'                    gives the respones of the \emph{ith} variable, ordered as in data_set_df. The columns denote the horizon.
 #'                    For example, if \emph{results_lin} contains the list with results, results_lin$irf_lin_mean[, , 1] returns a KXH matrix,
-#'                    where K is the number of variables and H the number of horizons. '1' is the variable which shocks, i.e. the
+#'                    where K is the number of variables and H the number of horizons. '1' is the variable shock variable, corresponding to the
 #'                    variable in the first column of \emph{data_set_df}.}
 #'
 #'\item{irf_lin_low}{A three 3D \link{array}() containing all lower confidence bands of the responses,

R/lp_nl.R

@@ -1,31 +1,32 @@
 #' @name lp_nl
 #' @title Compute nonlinear impulse responses
 #' @description Compute nonlinear impulse responses with local projections by Jordà (2005). The
-#' data are separated into two states with a smooth transition function as applied in Auerbach and Gorodnichenko (2012).
+#' data are separated into two states via a smooth transition function, applied in Auerbach and Gorodnichenko (2012).
 #'
 #' @param data_set_df A \link{data.frame}() containing all endogenous variables for the VAR. The column order
 #'                     is used for the Cholesky decomposition.
 #' @param lags_criterion NaN or character. NaN means that the number of lags
 #'         will be given at \emph{lags_nl} and \emph{lags_lin}. The lag length criteria are 'AICc', 'AIC' and 'BIC'.
-#' @param lags_lin NaN or integer. Number of lags for linear VAR to identify shock. NaN if lag length criterion is given.
-#' @param lags_nl NaN or integer. Number of lags for (nonlinear) VAR (if \emph{lags_criterion} = NaN). NaN if lag length criterion is given.
+#' @param lags_lin NaN or integer. NaN if lag length criterion is used.
+#'                                Integer for number of lags for linear VAR to identify shock.
+#' @param lags_nl NaN or integer. Number of lags for nonlinear VAR (if \emph{lags_criterion} = NaN). NaN if lag length criterion is given.
 #' @param max_lags NaN or integer. Maximum number of lags (if \emph{lags_criterion} = 'AICc', 'AIC', 'BIC'). NaN otherwise.
 #' @param trend Integer. Include no trend =  0 , include trend = 1, include trend and quadratic trend = 2.
 #' @param shock_type Integer. Standard deviation shock = 0, Unit shock = 1.
 #' @param confint Double. Width of confidence bands. 68\% = 1; 90\% = 1.65; 95\% = 1.96.
 #' @param hor Integer. Number of horizons for impulse responses.
 #' @param switching Vector. A column vector with the same length as \emph{data_set_df}. This series can either
 #'               be decomposed by the Hodrick-Prescott filter (see Auerbach and Gorodnichenko, 2013) or
-#'               directly plugged into the smooth transition function:
+#'               directly plugged into the following smooth transition function:
 #'               \deqn{ F_{z_t}) = \frac{exp(-\gamma z_t)}{1 + exp(-\gamma z_t)} }
 #'               Warning: \eqn{F_{z_t}} will be lagged in \link{create_nl_data} by one and then multiplied with the data.
 #'               If the variable shall not be lagged, the vector has to be given with a lead of one.
 #'               The data for the two regimes are: \cr
 #'               Regime 1 = (1-\eqn{F(z_{t-1})})*y_{t-p}, \cr
 #'               Regime 2 = \eqn{F(z_{t-1})}*y_{t-p}.
-#'@param gamma Double. Value of \eqn{\gamma} which is used in the transition function.
+#'@param gamma Double. Positive number which is used in the transition function.
 #'@param hp_filter Integer. No HP-filter = 0. Use HP-filter = 1.
-#'@param lambda Double. Value of \eqn{\lambda} for the Hodrick-Prescott filter if \emph{hp_filter} = 1.
+#'@param lambda Double. Value of \eqn{\lambda} for the Hodrick-Prescott filter if HP-filter is applied.
 #'
 #'
 #'
@@ -34,9 +35,9 @@
 #'
 #'\item{irf_s1_mean}{A three 3D \link{array}() containing all impulse responses for all endogenous variables of the first state.
 #'                    The last dimension denotes the shock variable. The row in each matrix
-#'                    denotes the respones of the \emph{ith} variable ordered as in \emph{data_set_df}. The columns are the horizons.
+#'                    denotes the respones of the \emph{ith} variable, ordered as in \emph{data_set_df}. The columns are the horizons.
 #'                    For example, if the results are saved in \emph{results_nl}, results_nl$irf_s1_mean[, , 1] returns a KXH matrix,
-#'                    where K is the number of variables and H the number of horizons. '1' is the variable which shocks, i.e. the
+#'                    where K is the number of variables and H the number of horizons. '1' is the shock variable, correspondong to the
 #'                    variable in the first column of \emph{data_set_df}.}
 #'
 #'\item{irf_s1_low}{A three 3D \link{array}() containing all lower confidence bands of the impulse responses, based on
@@ -47,7 +48,7 @@
 #'
 #'\item{irf_s2_mean}{A three 3D \link{array}() containing all impulse responses for all endogenous variables of the second state.
 #'                    The last dimension denotes the shock variable. The row in each matrix
-#'                    denotes the respones of the \emph{ith} variable as ordered in data_set_df. The columns denote the horizon.
+#'                    denotes the respones of the \emph{ith} variable, ordered as in data_set_df. The columns denote the horizon.
 #'                    For example, if the results are saved in \emph{results_nl}, results_nl$irf_s2_mean[, , 1] returns a KXH matrix,
 #'                    where K is the number of variables and H the number of horizons. '1' is the first shock variable corresponding to the
 #'                    variable in the first column of \emph{data_set_df}.}
@@ -60,7 +61,7 @@
 #'
 #'\item{specs}{A list with properties of \emph{data_set_df} for the plot function.}
 #'
-#'\item{fz}{A vector containing the values of the transition function F(z_{t-1})}
+#'\item{fz}{A vector containing the values of the transition function F(z_{t-1}).}
 #'
 #' @export
 #' @references

R/package.R

@@ -2,8 +2,8 @@
 #'
 #' Contains functions to estimate linear and nonlinear impulse
 #' responses based on local projections by Jordà (2005) <doi:10.1257/0002828053828518>.
-#' Nonlinear impulse responses are based on data separated by
-#' a transition function as in Auerbach and Gorodnichenko (2012) <doi:10.1257/pol.4.2.1>.
+#' To estimate nonlinear impulse responses the data are separated into two regimes by a
+#' a transition function, applied in Auerbach and Gorodnichenko (2012) <doi:10.1257/pol.4.2.1>.
 #' @docType package
 #' @author Philipp Adämmer
 #' @importFrom Rcpp evalCpp

R/plot_lin_irfs.R

@@ -1,5 +1,5 @@
 #' @name plot_lin_irfs
-#' @title Compute and display (linear) impulse responses
+#' @title Compute and display plots of linear impulse responses
 #' @description Compute and display (linear) impulse responses estimated with \link{lp_lin}.
 #' @param results_lin A list with 3D arrays estimated in \link{lp_lin}.
 #' @return A list with (gg-)plots for linear impulse responses

R/plot_nl_irfs.R

@@ -1,5 +1,5 @@
 #' @name plot_nl_irfs
-#' @title Compute and display (nonlinear) impulse responses
+#' @title Compute and display plots of nonlinear impulse responses
 #' @description Compute and display (nonlinear) impulse responses estimated with \link{lp_nl}().
 #' @param results_nl A list with 3D arrays estimated in \link{lp_nl}().
 #' @return A list with (gg-)plots for nonlinear impulse responses.

R/switching_series.R

@@ -1,7 +1,7 @@
 #' @name switching_series
-#' @title Function to estimate transition values
+#' @title Compute values of transition function to separate regimes
 #' @description Function to estimate transition values, which uses a smooth transition function as
-#' proposed by Auerbach & Gorodnichenko (2012). The time series used in the transition function
+#' used in Auerbach & Gorodnichenko (2012). The time series used in the transition function
 #' can be de-trended via the Hodrick-Prescott (see Auerbach and Gorodnichenko, 2013).
 #' @param switching_data A vector
 #' @param specs A list with specifications for \link{lp_nl}()

cran-comments.md

@@ -1,6 +1,20 @@
+## Release summary
+
+This is a first submission of the package to CRAN.
+
 ## Test environments
 * local x86_64-pc-linux-gnu (64-bit), R 3.4.4
 * ubuntu 12.04 (on travis-ci), R 3.1.2
 * win-builder (devel and release)
 
+## R CMD check results
+
+0 errors ✔ | 0 warnings ✔ | 1 note ✖
 
+checking CRAN incoming feasibility ... NOTE
+  Maintainer: ‘Philipp Adämmer <adaemmer@hsu-hh.de>’
+
+  Possibly mis-spelled words in DESCRIPTION:
+    Auerbach (16:44)
+    Gorodnichenko (16:57)
+    Jordà (4:12, 14:54)

docs/README_docs.Rmd

@@ -116,15 +116,15 @@ $$ $$
 Estimate nonlinear impulse responses with package function *lp_nl*. Note that the VAR's data set (*data_set_df*) has to be 
 a data.frame(). 
 ```{r, message = FALSE }  
-results_nl <- lp_nl(data_set_df, lags_lin         = 4L,  # Number of lags for (reduced) linear VAR 
+results_nl <- lp_nl(data_set_df, lags_lin       = 4L,  # Number of lags for (reduced) linear VAR 
                                  lags_nl        = 4L,  # Number of lags 
                                  lags_criterion = NaN, # Lag length criterion (AICc, AIC or BIC) or NaN
                                  max_lags       = NaN, # If lags_criterion is chosen, maximum number of lags  
                                  trend          = 1L,  # 0 = no trend, 1 = trend, 2 = trend and quadratic trend
                                  shock_type     = 0L,  # 0 = standard deviation shock, 1 = unit shock
                                  confint        = 1.67, # Width of confidence bands: 1 = 68%, 1.67 = 90%, 
                                                         # 1.96 = 95%
-                                 hor            = 24,  # Length of horizon
+                                 hor            = 24,   # Length of horizon
                       
                                  switching      = data_set_df$Infl, # Inflation rate
                                  hp_filter      = 1,    # 0 = Do not use HP-filter, 1 = use HP-filter   

docs/README_docs.html

@@ -182,15 +182,15 @@ <h2>Example: Nonlinear impulse responses</h2>
 \]</span> with <span class="math inline">\(h = 1,..., H.\)</span></p>
 <p><span class="math display">\[ \]</span></p>
 <p>Estimate nonlinear impulse responses with package function <em>lp_nl</em>. Note that the VAR’s data set (<em>data_set_df</em>) has to be a data.frame().</p>
-<pre class="r"><code>results_nl &lt;- lp_nl(data_set_df, lags_lin         = 4L,  # Number of lags for (reduced) linear VAR 
+<pre class="r"><code>results_nl &lt;- lp_nl(data_set_df, lags_lin       = 4L,  # Number of lags for (reduced) linear VAR 
                                  lags_nl        = 4L,  # Number of lags 
                                  lags_criterion = NaN, # Lag length criterion (AICc, AIC or BIC) or NaN
                                  max_lags       = NaN, # If lags_criterion is chosen, maximum number of lags  
                                  trend          = 1L,  # 0 = no trend, 1 = trend, 2 = trend and quadratic trend
                                  shock_type     = 0L,  # 0 = standard deviation shock, 1 = unit shock
                                  confint        = 1.67, # Width of confidence bands: 1 = 68%, 1.67 = 90%, 
                                                         # 1.96 = 95%
-                                 hor            = 24,  # Length of horizon
+                                 hor            = 24,   # Length of horizon
 
                                  switching      = data_set_df$Infl, # Inflation rate
                                  hp_filter      = 1,    # 0 = Do not use HP-filter, 1 = use HP-filter