version 0.0-2

DESCRIPTION

@@ -1,15 +1,15 @@
 Package: fastM
 Type: Package
 Title: Fast Computation of Multivariate M-estimators
-Version: 0.0-1
-Date: 2014-02-27
-Author: Lutz Duembgen, Klaus Nordhausen, Heike Schumacher 
+Version: 0.0-2
+Date: 2014-04-12
+Author: Lutz Duembgen, Klaus Nordhausen, Heike Schuhmacher 
 Maintainer: Klaus Nordhausen <klaus.nordhausen@utu.fi>
 Imports: Rcpp (>= 0.11.0)
 LinkingTo: Rcpp, RcppArmadillo
 Description: The package implements the new algorithm for fast computation of M-scatter matrices using a partial Newton-Raphson procedure for several estimators. 
 License: GPL (>= 2)
-Packaged: 2014-02-27 07:07:33 UTC; klanor
+Packaged: 2014-04-12 08:32:18 UTC; klanor
 NeedsCompilation: yes
 Repository: CRAN
-Date/Publication: 2014-02-27 11:49:29
+Date/Publication: 2014-04-12 23:26:01

MD5

@@ -1,4 +1,4 @@
-987fb6e2876aef4aa504dac42b2e943a *DESCRIPTION
+031e170e828fed2ddfe5439bb5ec6441 *DESCRIPTION
 9f40234a4d7001dfd8a20c879b6d107d *NAMESPACE
 0a3836f37772c618d6ff3bea4b70268c *R/Duembgen.R
 cad292203fc052d769d2645a20c67bef *R/MVTMLE0wrap.R
@@ -12,13 +12,13 @@ ae2a4230842f203ee74c161269c3fa36 *R/MVTMLEsymm2wrap.R
 2a2709d3f74d564ad12faa17547c2e38 *R/TYLERsymm1wrap.R
 747a05ed8acc276920f65d6adee9e32c *R/TYLERsymm2wrap.R
 29867a32cd93313bb8f143613cbf50e3 *R/TYLERwrap.R
-1d3e77be673d3f644cb58f8c3d155385 *inst/ChangeLog
-684b9a52c59c0ae3b2f262e7b9c2c3d2 *man/DUEMBGENshape.Rd
-ef9dffc3af21c4ed23d1bef6ceb8f422 *man/MVTMLE.Rd
-99e82bffc3a2be41fa404ff092e146d9 *man/MVTMLE0r.Rd
-dc7d309a4d1e9823dd6ccf9a3f561fdf *man/MVTMLEsymm.Rd
-8e8f578469ffd23fe5320bd292836757 *man/TYLERshape.Rd
-3d1e3b401dc9b6c63b424b8625c99406 *man/fastM-package.Rd
+6d27758b928c944363796241d67cb387 *inst/ChangeLog
+5c5487c2ea429f567bdc3fa28c72f256 *man/DUEMBGENshape.Rd
+a8c1fb77f61a2a1fa854d3a46870bc73 *man/MVTMLE.Rd
+db2201eeb346d30706b269b7e173a444 *man/MVTMLE0r.Rd
+929c3dea97588a4f389f546bbf5104e6 *man/MVTMLEsymm.Rd
+6acc35dc7f979f99230968e6a8eb3810 *man/TYLERshape.Rd
+69522be81338a160065643cd8ffbe745 *man/fastM-package.Rd
 d4930fc21ea66dfe3e1bf3529823b741 *src/MVTMLE.cpp
 12e73e0c1a75e09ab9381e067e5e6f40 *src/MVTMLE.h
 95a5e0dc5dbdec9c251c142cb4cad286 *src/Makevars

inst/ChangeLog

@@ -1,4 +1,7 @@
 LOG of the CHANGES in the package fastM
 
 Version 0.0-1:
-* First build
+* First build
+
+Version 0.0-2:
+* Improving help files

man/DUEMBGENshape.Rd

@@ -15,9 +15,9 @@ DUEMBGENshape(X, nmax = 500, eps = 1e-06, maxiter = 100, perm = FALSE)
   \item{X}{numeric data matrix or dataframe. Missing values are not allowed.}
   \item{nmax}{integer, if the sample size n (number of rows of \code{X}) is smaller than \code{nmax}, then all n(n-1)/2 pairwise differences will be computed
 and used in the algorithm. If n is larger, then the algorithm avoids storing all the pairwise differences and is more memory efficient.}
-  \item{eps}{convergence tolerance.}
+  \item{eps}{convergence tolerance, which means that the algorithm stops when the Frobenius norm of the gradient is smaller than eps.}
   \item{maxiter}{maximum number of iterations.}
-  \item{perm}{logical. If TRUE the rows of the \code{X} will be randomly permuted before starting the computations. See details.}
+  \item{perm}{logical. If TRUE the rows of \code{X} will be randomly permuted before starting the computations. See details.}
 }
 \details{
 The estimate is based on the new fast algorithm described in Duembgen et al. (2014).
@@ -29,7 +29,7 @@ To get a good initial value for the algorithm, the estimator is first computed b
 successive observations. Therefore the order of the rows of \code{X} is supposed to be random. If this is not the case, the data
 should be first permuted using the argument \code{perm}.
 
-In case \code{maxiter} is reached before convergence, the estimate at that iteration is returned and a warning given.
+In case \code{maxiter} is reached before convergence, the estimate at that iteration is returned and a warning is given.
 }
 \value{
   A list containing:
@@ -39,7 +39,7 @@ In case \code{maxiter} is reached before convergence, the estimate at that itera
 \references{
 \cite{Duembgen, L. (1998), On Tyler's M-functional of scatter in high dimension, \emph{Annals of Institute of Statistical Mathematics}, \bold{50},  471--491. }
 
-\cite{Duembgen, L., Nordhausen, K. and Schumacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as	arXiv:1312.6489.}
+\cite{Duembgen, L., Nordhausen, K. and Schuhmacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as	arXiv:1312.6489.}
 }
 \author{
 Lutz Duembgen and Klaus Nordhausen

man/MVTMLE.Rd

@@ -5,7 +5,7 @@
 M-estimator of Location and Scatter Using Weights Coming From the Multivariate t-distribution
 }
 \description{
-The algorithm of this function is a based on a partial Newton approach and should be faster than the traditional fixed-point algorithm.
+The algorithm of this function is based on a partial Newton approach and should be faster than the traditional fixed-point algorithm.
 If the data follows a multivariate t-distribution with the correctly specified degrees of freedom this function gives the maximum likelihood estimate
 of location and scatter.
 }
@@ -16,15 +16,15 @@ MVTMLE(X, nu = 1, location = TRUE, eps = 1e-06, maxiter = 100)
 \arguments{
   \item{X}{numeric data matrix or dataframe. Missing values are not allowed.}
   \item{nu}{assumed degrees of freedom of the t-distribution. Default is '1' which corresponds to the Cauchy distribution.}
-  \item{location}{logical or numeric. If FALSE, it is assume that the scatter should be computed wrt to the origin. If TRUE the location will be estimated and if it is a numeric
-  vector it will be computed wrt to this vector}
-  \item{eps}{convergence tolerance.}
+  \item{location}{logical or numeric. If FALSE, it is assumed that the scatter should be computed wrt to the origin. If TRUE the location will be estimated and if it is a numeric
+  vector it will be computed wrt to this vector.}
+  \item{eps}{convergence tolerance, which means that the algorithm stops when the Frobenius norm of the gradient is smaller than eps.}
   \item{maxiter}{maximum number of iterations.}
 }
 \details{
 The assumed degree of freedom nu must be at least 1 when the location and scatter should be estimated. If only the scatter is to be estimated, then it needs to be larger than zero only. 
 
-In case \code{maxiter} is reached before convergence, the estimate at that iteration is returned and a warning given.
+In case \code{maxiter} is reached before convergence, the estimate at that iteration is returned and a warning is given.
 }
 \value{
  A list containing:
@@ -35,7 +35,7 @@ In case \code{maxiter} is reached before convergence, the estimate at that itera
 \references{
 \cite{Kent, J.T., Tyler, D.E. and Vardi, Y. (1994),  A curious likelihood identity for the multivariate t-distribution, \emph{Communications in Statistics, Theory and Methods}, \bold{23},  441--453.}
 
-\cite{Duembgen, L., Nordhausen, K. and Schumacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as arXiv:1312.6489.}
+\cite{Duembgen, L., Nordhausen, K. and Schuhmacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as arXiv:1312.6489.}
 }
 \author{
 Lutz Duembgen and Klaus Nordhausen

man/MVTMLE0r.Rd

@@ -23,15 +23,15 @@ MVTMLE0r_CG(X, nu = 0, delta = 1e-06, steps = FALSE)
 \arguments{
   \item{X}{numeric data matrix or dataframe. Missing values are not allowed.}
   \item{nu}{assumed degrees of freedom of the t-distribution. Must be 0 or larger. Default is '0' which corresponds to Tyler's shape matrix.}
-  \item{delta}{convergence tolerance.}
+  \item{delta}{convergence tolerance, which means that the algorithms stop when the Frobenius norm of the gradient is smaller than delta.}
   \item{prewhitened}{logical. Is the data prewhitened or not.}
   \item{steps}{logial. If TRUE intermediate results are printed on the console.}
 }
 \details{
 All functions are implemented in R and their purpose is only for demonstration of the differences of the different algorithms.
 The function \code{MVTMLE0r} uses the recommended partial Newton approach as implemented also in (\code{\link{MVTMLE}} and \code{\link{TYLERshape}}).
 \code{MVTMLE0r_FP} and \code{MVTMLE0r_FP0} are fixed-point algorithms where \code{MVTMLE0r_FP} iterates the fixed point equation with
-'iterative whitening' of the data. The function \code{MVTMLE0r_G} uses a gradiant approach and \code{MVTMLE0r_CG} a conjugate gradiant approach.
+'iterative whitening' of the data. The function \code{MVTMLE0r_G} uses a gradient approach and \code{MVTMLE0r_CG} a conjugate gradient approach.
 Note that \code{MVTMLE0r_CG} does not check if the 'next' step is really an improvement and that all functions compute the scatter wrt to the origin. 
 
 All functions have a hard coded maximum number of iterations of 1000. If that is reached the functions returns the final estimate, however without a warning.
@@ -46,7 +46,7 @@ For general purposes we recommend the functions \code{\link{MVTMLE}} and \code{\
   \item{iter}{Number of iterations of the algorithm.}
 }
 \references{
-\cite{Duembgen, L., Nordhausen, K. and Schumacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as arXiv:1312.6489.}
+\cite{Duembgen, L., Nordhausen, K. and Schuhmacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as arXiv:1312.6489.}
 }
 \author{
 Lutz Duembgen and Klaus Nordhausen

man/MVTMLEsymm.Rd

@@ -19,24 +19,24 @@ MVTMLEsymm(X, nu = 1, nmax = 500, eps = 1e-06, maxiter = 100, perm = FALSE)
 \item{nu}{assumed degrees of freedom of the t-distribution, must be larger than 0. Default is '1'.}
 \item{nmax}{integer, if the sample size n (number of rows of \code{X}) is smaller than \code{nmax}, then all n(n-1)/2 pairwise differences will be computed
 and used in the algorithm. If n is larger, then the algorithm avoids storing all the pairwise differences and is more memory efficient.}
-\item{eps}{convergence tolerance.}
+\item{eps}{convergence tolerance, which means that the algorithm stops when the Frobenius norm of the gradient is smaller than eps.}
 \item{maxiter}{maximum number of iterations.}
-\item{perm}{logical. If TRUE the rows of the \code{X} will be randomly permuted before starting the computations. See details.}
+\item{perm}{logical. If TRUE the rows of \code{X} will be randomly permuted before starting the computations. See details.}
 }
 \details{
 To get a good initial value for the algorithm, the estimator is first computed based on the pairwise differences of 
 successive observations. Therefore the order of the rows of \code{X} is supposed to be random. If this is not the case, the data
 should be first permuted using the argument \code{perm}.
 
-In case \code{maxiter} is reached before convergence, the estimate at that iteration is returned and a warning given.
+In case \code{maxiter} is reached before convergence, the estimate at that iteration is returned and a warning is given.
 }
 \value{
  A list containing:
   \item{Sigma}{Estimated scatter matrix.}
   \item{iter}{Number of iterations of the algorithm.}
 }
 \references{
-\cite{Duembgen, L., Nordhausen, K. and Schumacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as arXiv:1312.6489.}
+\cite{Duembgen, L., Nordhausen, K. and Schuhmacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as arXiv:1312.6489.}
 }
 \author{
 Lutz Duembgen and Klaus Nordhausen

man/TYLERshape.Rd

@@ -10,19 +10,20 @@
 
 \arguments{
   \item{X}{numeric data matrix or dataframe. Missing values are not allowed.}
-  \item{location}{logical. If TRUE it will be estimated with respect to the mean and otherwise with respect to the origin.}
-  \item{eps}{convergence tolerance.}
+  \item{location}{logical or numeric. If FALSE, it is assumed that the scatter should be computed wrt to the origin. If TRUE the location will be estimated as the mean vector and if it is a numeric
+  vector it will be computed wrt to the given vector.}
+  \item{eps}{convergence tolerance, which means that the algorithm stops when the Frobenius norm of the gradient is smaller than eps.}
   \item{maxiter}{maximum number of iterations.}
 }
 
 \details{
-The estimate is based on the new fast algorithm described in Duembgen et al.~(2014).
+The estimate is based on the new fast algorithm described in Duembgen et al. (2014).
 Note that Tyler's shape matrix is standardized such that it has determinant 1.
 
 The function does not check if there are observations equal to the mean (if \code{location=TRUE}), to the provided location vector or to the origin (if \code{location=FALSE}).
 In these cases the function will fail.
 
-In case \code{maxiter} is reached before convergence, the estimate at that iteration is returned and a warning given.
+In case \code{maxiter} is reached before convergence, the estimate at that iteration is returned and a warning is given.
 }
 \value{
   A list containing:
@@ -33,7 +34,7 @@ In case \code{maxiter} is reached before convergence, the estimate at that itera
 \references{
 \cite{Tyler, D.E. (1987), A distribution-free M-estimator of scatter, \emph{Annals of Statistics}, \bold{15},  234--251. } 
 
-\cite{Duembgen, L., Nordhausen, K. and Schumacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as	arXiv:1312.6489.}
+\cite{Duembgen, L., Nordhausen, K. and Schuhmacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as	arXiv:1312.6489.}
 }
 \author{
 Lutz Duembgen and Klaus Nordhausen

man/fastM-package.Rd

@@ -12,8 +12,8 @@ The package implements the new algorithm for fast computation of M-scatter matri
 \tabular{ll}{
 Package: \tab fastM\cr
 Type: \tab Package\cr
-Version: \tab 0.1\cr
-Date: \tab 2014-02-27\cr
+Version: \tab 0.2\cr
+Date: \tab 2014-04-12\cr
 License: \tab GPL (>= 2)
 }
 Multivariate M-estimators are usually computed using a fixed-point algorithm. As recently shown in Duembgen et al. (2014) 
@@ -25,15 +25,15 @@ For the symmetrized M-estimators we work with incomplete U-statistics to acceler
 
 
 \author{
-Lutz Duembgen, Klaus Nordhausen, Heike Schumacher
+Lutz Duembgen, Klaus Nordhausen, Heike Schuhmacher
 
 Maintainer: Klaus Nordhausen <klaus.nordhausen@utu.fi> 
 
 }
 \references{
 \cite{Duembgen, L. (1998), On Tyler's M-functional of scatter in high dimension, \emph{Annals of Institute of Statistical Mathematics}, \bold{50},  471--491. }
 
-\cite{Duembgen, L., Nordhausen, K. and Schumacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as	arXiv:1312.6489.}
+\cite{Duembgen, L., Nordhausen, K. and Schuhmacher, H. (2014),  New algorithms for M-estimation of multivariate location and scatter, preprint available as	arXiv:1312.6489.}
 
 \cite{Kent, J.T., Tyler, D.E. and Vardi, Y. (1994),  A curious likelihood identity for the multivariate t-distribution, \emph{Communications in Statistics, Theory and Methods}, \bold{23},  441--453.}