Final touches before submitting to CRAN.

Author: dppalomar

DESCRIPTION

@@ -1,12 +1,13 @@
 Package: fitHeavyTail
 Title: Mean and Covariance Matrix Estimation under Heavy Tails
-Version: 0.1.2.9000
-Date: 2020-1-7
+Version: 0.1.3
+Date: 2022-4-14
 Description: Robust estimation methods for the mean vector, scatter matrix,
     and covariance matrix (if it exists) from data (possibly containing NAs) 
     under multivariate heavy-tailed distributions such as angular Gaussian 
     (via Tyler's method), Cauchy, and Student's t distributions. Additionally, 
-    a factor model structure can be specified for the covariance matrix.
+    a factor model structure can be specified for the covariance matrix. The
+    latest revision also includes the multivariate skewed t distribution.
     The package is based on the papers: Sun, Babu, and Palomar (2014),
     Sun, Babu, and Palomar (2015), Liu and Rubin (1995), and 
     Zhou, Liu, Kumar, and Palomar (2019).
@@ -24,7 +25,6 @@ URL: https://CRAN.R-project.org/package=fitHeavyTail,
 BugReports: https://github.com/dppalomar/fitHeavyTail/issues
 License: GPL-3
 Encoding: UTF-8
-LazyData: true
 RoxygenNote: 7.1.1
 Depends:
 Imports:

NAMESPACE

@@ -11,6 +11,8 @@ importFrom(mvtnorm,dmvt)
 importFrom(numDeriv,grad)
 importFrom(stats,cov)
 importFrom(stats,median)
+importFrom(stats,na.omit)
 importFrom(stats,optimize)
+importFrom(stats,uniroot)
 importFrom(stats,var)
 importFrom(utils,tail)

NEWS.md

@@ -1,4 +1,4 @@
-## Changes in fitHeavyTail version 0.1.2.9000
+## Changes in fitHeavyTail version 0.1.3 (2022-4-14)
 
 * New method for skewed t distributions: fit_mvst()
 

R/fit_mvt.R

@@ -57,6 +57,7 @@
 #'                         \item{\code{psi}: default is
 #'                                          \code{diag(initial$cov - initial$B \%*\% t(initial$B)).}}}
 #' @param optimize_mu Boolean indicating whether to optimize \code{mu} (default is \code{TRUE}).
+#' @param weights Optional weights for each of the observations (the length should be equal to the number of rows of X).
 #' @param factors Integer indicating number of factors (default is \code{ncol(X)}, so no factor model assumption).
 #' @param max_iter Integer indicating the maximum number of iterations for the iterative estimation
 #'                 method (default is \code{100}).
@@ -104,6 +105,9 @@
 #' Rui Zhou, Junyan Liu, Sandeep Kumar, and Daniel P. Palomar, "Robust factor analysis parameter estimation,"
 #' Lecture Notes in Computer Science (LNCS), 2019. <https://arxiv.org/abs/1909.12530>
 #'
+#' Esa Ollila, Daniel P. Palomar, and Frédéric Pascal, "Shrinking the Eigenvalues of M-estimators of Covariance Matrix,"
+#' IEEE Trans. on Signal Processing, vol. 69, pp. 256-269, Jan. 2021.
+#'
 #' @examples
 #' library(mvtnorm)       # to generate heavy-tailed data
 #' library(fitHeavyTail)
@@ -115,7 +119,7 @@
 #' options(nu_min = 4.01)
 #' fit_mvt(X, nu = "iterative")
 #'
-#' @importFrom stats optimize
+#' @importFrom stats optimize na.omit uniroot
 #' @export
 fit_mvt <- function(X, na_rm = TRUE,
                     nu = c("kurtosis", "MLE-diag", "MLE-diag-resampled", "iterative"),
@@ -330,7 +334,7 @@ fit_mvt <- function(X, na_rm = TRUE,
                        theta <- (1 - N/T) * sum(r2i)/T/N
                        # correction with sigma
                        T_ <- 10000
-                       X_ <- mvtnorm::rmvt(n = T_, delta = rep(0, N), sigma = diag(N), df = nu_true)  # heavy-tailed data
+                       X_ <- mvtnorm::rmvt(n = T_, delta = rep(0, N), sigma = diag(N), df = nu)  # heavy-tailed data
                        r2 <- rowSums(X_^2)
                        u <- (N + nu)/(nu + r2)
                        r2i <- r2/(1 - r2*u/T_)

R_buildignore/developer_commands.R

@@ -39,6 +39,6 @@ devtools::build()
 #devtools::revdep(pkg = "fitHeavyTail")  # to check reverse dependencies
 #devtools::check_win_release()  #to check under windows
 #R CMD build .  # this is to generate tarball
-#R CMD check fitHeavyTail_0.1.2.9000.tar.gz --as-cran --run-donttest  # this is before submission to CRAN
-#R CMD install fitHeavyTail_0.1.2.9000.tar.gz
+#R CMD check fitHeavyTail_0.1.3.tar.gz --as-cran --run-donttest  # this is before submission to CRAN
+#R CMD install fitHeavyTail_0.1.3.tar.gz
 #submit the tarball directly via the webform: https://cran.r-project.org/submit.html

cran-comments.md

@@ -1,9 +1,9 @@
-This is version 0.1.2 (2020-1-7) in CRAN (package also available in GitHub).
+This is version 0.1.3 (2022-4-14) in CRAN (package also available in GitHub).
 
 ---
 
 ## Test environments
-* OS X, R 3.6.0
+* OS X, R 4.1.2
 * Windows at https://win-builder.r-project.org/
 
 ## R CMD check results

inst/CITATION

@@ -3,14 +3,15 @@ citHeader("To cite 'fitHeavyTail' in publications, please use:")
 citEntry(entry = "Manual",
          title        = "{fitHeavyTail: Mean and Covariance Matrix Estimation under Heavy Tails}",
          author       = personList(as.person("D. P. Palomar"),
-                                   as.person("Rui ZHOU")),
-         note         = "R package version 0.1.2",
-         year         = "2020",
+                                   as.person("Rui Zhou"),
+                                   as.person("Xiwen Wang")),
+         note         = "R package version 0.1.3",
+         year         = "2022",
          url          = "https://CRAN.R-project.org/package=fitHeavyTail",
          textVersion  =
-           paste("D. P. Palomar and Rui Zhou (2020).",
+           paste("D. P. Palomar, Rui Zhou, and Xiwen Wang (2022).",
                  "fitHeavyTail: Mean and Covariance Matrix Estimation under Heavy Tails.",
-                 "R package version 0.1.2.",
+                 "R package version 0.1.3.",
                  "https://CRAN.R-project.org/package=fitHeavyTail")
 )
 
@@ -51,3 +52,22 @@ citEntry(entry = "Article",
                  "vol. 63, no. 12, pp. 3096-3109.",
                  "https://doi.org/10.1109/TSP.2015.2417513")
 )
+
+citEntry(entry = "Article",
+         title        = "Shrinking the Eigenvalues of M-estimators of Covariance Matrix",
+         author       = personList(as.person("E. Ollila"),
+                                   as.person("D. P. Palomar"),
+                                   as.person("F. Pascal")),
+         journal      = "IEEE Transactions on Signal Processing",
+         volume       = "69",
+         pages        = "256--269",
+         year         = "2021",
+         url          = "https://doi.org/10.1109/TSP.2020.3043952",
+         textVersion  =
+           paste("E. Ollila, D. P. Palomar, and F. Pascal (2021).",
+                 "Shrinking the Eigenvalues of M-estimators of Covariance Matrix.",
+                 "IEEE Transactions on Signal Processing,",
+                 "vol. 69, pp. 256-269.",
+                 "https://doi.org/10.1109/TSP.2020.3043952")
+)
+

man/fit_mvt.Rd

@@ -39,8 +39,10 @@ test_that("default mode works", {
   load("fitted_mvst_check.RData")
 
   mvst_model <- fit_mvst(X, ptol = 1e-4)
+
   expect_equal(mvst_model[c("mu", "gamma", "scatter", "nu", "mean", "cov", "converged", "num_iterations")],
-               mvst_model_check[c("mu", "gamma", "scatter", "nu", "mean", "cov", "converged", "num_iterations")])  #, tolerance = 0.1)
+               mvst_model_check[c("mu", "gamma", "scatter", "nu", "mean", "cov", "converged", "num_iterations")],
+               tolerance = 1e-5)
 
   # test for xts
   mvst_model <- fit_mvst(X)

tests/testthat/test-fit_mvst.R

@@ -22,7 +22,7 @@ test_that("error control works", {
                "Cannot deal with T <= N (after removing NAs), too few samples.", fixed = TRUE)
   expect_error(fit_mvt(X = X, max_iter = -1), "\"max_iter\" must be greater than 1.")
   expect_error(fit_mvt(X = X, factors = -1), "\"factors\" must be no less than 1 and no more than column number of \"X\".")
-  expect_error(fit_mvt(X = X, nu = "lala"), "'arg' should be one of “kurtosis”, “MLE-diag”, “MLE-diag-resampled”, “iterative”")
+  expect_error(fit_mvt(X = X, nu = "lala"), "'arg' should be one of \"kurtosis\", \"MLE-diag\", \"MLE-diag-resampled\", \"iterative\"")
   expect_error(fit_mvt(X = X, nu = 1), "Non-valid value for nu.")
 })