2
0
Fork 0

fix: typos (in Doc comments)

This commit is contained in:
Daniel Kapla 2019-10-22 10:33:41 +02:00
parent 92b1a49d2b
commit 10ae55bd81
12 changed files with 26 additions and 26 deletions

View File

@ -43,7 +43,7 @@
#' supplied. #' supplied.
#' @param method specifies the CVE method variation as one of #' @param method specifies the CVE method variation as one of
#' \itemize{ #' \itemize{
#' \item "simple" exact implementation as describet in the paper listed #' \item "simple" exact implementation as described in the paper listed
#' below. #' below.
#' \item "weighted" variation with addaptive weighting of slices. #' \item "weighted" variation with addaptive weighting of slices.
#' } #' }
@ -63,7 +63,7 @@
#' dr <- cve(Y ~ X) #' dr <- cve(Y ~ X)
#' round(dr[[2]]$B, 1) #' round(dr[[2]]$B, 1)
#' #'
#' @seealso For a detailed description of the formula parameter see #' @seealso For a detailed description of \code{formula} see
#' [\code{\link{formula}}]. #' [\code{\link{formula}}].
#' @export #' @export
cve <- function(formula, data, method = "simple", max.dim = 10L, ...) { cve <- function(formula, data, method = "simple", max.dim = 10L, ...) {
@ -90,16 +90,15 @@ cve <- function(formula, data, method = "simple", max.dim = 10L, ...) {
#' @param nObs parameter for choosing bandwidth \code{h} using #' @param nObs parameter for choosing bandwidth \code{h} using
#' \code{\link{estimate.bandwidth}} (ignored if \code{h} is supplied). #' \code{\link{estimate.bandwidth}} (ignored if \code{h} is supplied).
#' @param X data matrix with samples in its rows. #' @param X data matrix with samples in its rows.
#' @param Y Responces (1 dimensional). #' @param Y Responses (1 dimensional).
#' @param k Dimension of lower dimensional projection, if given only the #' @param k Dimension of lower dimensional projection, if \code{k} is given only the specified dimension \code{B} matrix is estimated.
#' specified dimension is estimated.
#' @param min.dim lower bounds for \code{k}, (ignored if \code{k} is supplied). #' @param min.dim lower bounds for \code{k}, (ignored if \code{k} is supplied).
#' @param max.dim upper bounds for \code{k}, (ignored if \code{k} is supplied). #' @param max.dim upper bounds for \code{k}, (ignored if \code{k} is supplied).
#' @param tau Initial step-size. #' @param tau Initial step-size.
#' @param tol Tolerance for break condition. #' @param tol Tolerance for break condition.
#' @param epochs maximum number of optimization steps. #' @param epochs maximum number of optimization steps.
#' @param attempts number of arbitrary different starting points. #' @param attempts number of arbitrary different starting points.
#' @param logger a logger function (only for addvanced user). #' @param logger a logger function (only for advanced user, significantly slows down the computation).
#' @rdname cve #' @rdname cve
#' @export #' @export
cve.call <- function(X, Y, method = "simple", cve.call <- function(X, Y, method = "simple",
@ -235,7 +234,7 @@ cve.call <- function(X, Y, method = "simple",
return(dr) return(dr)
} }
#' Loss distribution kink plot. #' Loss distribution elbow plot.
#' #'
#' @param x Object of class \code{"cve"} (result of [\code{\link{cve}}]). #' @param x Object of class \code{"cve"} (result of [\code{\link{cve}}]).
#' @param ... Pass through parameters to [\code{\link{plot}}] and #' @param ... Pass through parameters to [\code{\link{plot}}] and
@ -245,6 +244,7 @@ cve.call <- function(X, Y, method = "simple",
#' as well as \code{\link{plot}}, the standard plot utility. #' as well as \code{\link{plot}}, the standard plot utility.
#' @importFrom graphics plot lines points #' @importFrom graphics plot lines points
#' @method plot cve #' @method plot cve
#' Boxplots of the loss from \code{min.dim} to \code{max.dim} \code{k} values.
#' @export #' @export
plot.cve <- function(x, ...) { plot.cve <- function(x, ...) {
L <- c() L <- c()
@ -256,7 +256,7 @@ plot.cve <- function(x, ...) {
} }
} }
L <- matrix(L, ncol = length(k)) / var(x$Y) L <- matrix(L, ncol = length(k)) / var(x$Y)
boxplot(L, main = "Kink plot", boxplot(L, main = "elbow plot",
xlab = "SDR dimension", xlab = "SDR dimension",
ylab = "Sample loss distribution", ylab = "Sample loss distribution",
names = k) names = k)

View File

@ -1,7 +1,7 @@
#' Generates test datasets. #' Generates test datasets.
#' #'
#' Provides sample datasets. There are 5 different datasets named #' Provides sample datasets. There are 5 different datasets named
#' M1, M2, M3, M4 and M5 describet in the paper references below. #' M1, M2, M3, M4 and M5 described in the paper references below.
#' The general model is given by: #' The general model is given by:
#' \deqn{Y ~ g(B'X) + \epsilon} #' \deqn{Y ~ g(B'X) + \epsilon}
#' #'

View File

@ -1,16 +1,16 @@
#' Bandwidth estimation for CVE. #' Bandwidth estimation for CVE.
#' #'
#' Estimates a propper bandwidth \code{h} according #' Estimates a bandwidth \code{h} according
#' \deqn{% #' \deqn{%
#' h = \chi_{k}^{-1}\left(\frac{nObs - 1}{n-1}\right)\frac{2 tr(\Sigma)}{p}}{% #' h = \chi_{k}^{-1}\left(\frac{nObs - 1}{n-1}\right)\frac{2 tr(\Sigma)}{p}}{%
#' h = qchisq( (nObs - 1)/(n - 1), k ) * (2 tr(\Sigma) / p)} #' h = qchisq( (nObs - 1)/(n - 1), k ) * (2 tr(\Sigma) / p)}
#' with \eqn{n} the number of sample and \eqn{p} its dimension #' with \eqn{n} the sample size, \eqn{p} its dimension
#' (\code{n <- nrow(X); p <- ncol(X)}) and the covariance-matrix \eqn{\Sigma} #' (\code{n <- nrow(X); p <- ncol(X)}) and the covariance-matrix \eqn{\Sigma}
#' which is given by the standard maximum likelihood estimate. #' which is \code{(n-1)/n} times the sample covariance estimate.
#' #'
#' @param nObs Expected number of points in a slice, see paper.
#' @param X data matrix with samples in its rows. #' @param X data matrix with samples in its rows.
#' @param k Dimension of lower dimensional projection. #' @param k Dimension of lower dimensional projection.
#' @param nObs number of points in a slice, see \eqn{nObs} in CVE paper.
#' #'
#' @seealso [\code{\link{qchisq}}] #' @seealso [\code{\link{qchisq}}]
#' @export #' @export

View File

@ -1,8 +1,8 @@
#' Samples uniform from the Stiefl Manifold. #' Draws a sample from the invariant measure on the Stiefel manifold \eqn{S(p, q)}.
#' #'
#' @param p row dim. #' @param p row dimension
#' @param q col dim. #' @param q col dimension
#' @return `(p, q)` semi-orthogonal matrix #' @return \code{p} times \code{q} semi-orthogonal matrix.
#' @examples #' @examples
#' V <- rStiefel(6, 4) #' V <- rStiefel(6, 4)
#' @export #' @export

View File

@ -20,7 +20,7 @@ supplied.}
\item{method}{specifies the CVE method variation as one of \item{method}{specifies the CVE method variation as one of
\itemize{ \itemize{
\item "simple" exact implementation as describet in the paper listed \item "simple" exact implementation as described in the paper listed
below. below.
\item "weighted" variation with addaptive weighting of slices. \item "weighted" variation with addaptive weighting of slices.
}} }}

View File

@ -28,7 +28,7 @@ List with elements
} }
\description{ \description{
Provides sample datasets. There are 5 different datasets named Provides sample datasets. There are 5 different datasets named
M1, M2, M3, M4 and M5 describet in the paper references below. M1, M2, M3, M4 and M5 described in the paper references below.
The general model is given by: The general model is given by:
\deqn{Y ~ g(B'X) + \epsilon} \deqn{Y ~ g(B'X) + \epsilon}
} }

View File

@ -2,7 +2,7 @@
% Please edit documentation in R/CVE.R % Please edit documentation in R/CVE.R
\name{plot.cve} \name{plot.cve}
\alias{plot.cve} \alias{plot.cve}
\title{Creates a kink plot of the sample loss distribution over SDR dimensions.} \title{Loss distribution elbow plot.}
\usage{ \usage{
\method{plot}{cve}(x, ...) \method{plot}{cve}(x, ...)
} }
@ -13,7 +13,7 @@
[\code{\link{lines}}]} [\code{\link{lines}}]}
} }
\description{ \description{
Creates a kink plot of the sample loss distribution over SDR dimensions. Loss distribution elbow plot.
} }
\seealso{ \seealso{
see \code{\link{par}} for graphical parameters to pass through see \code{\link{par}} for graphical parameters to pass through

View File

@ -76,7 +76,7 @@ cve <- function(formula, data, method = "simple", max.dim = 10L, ...) {
return(dr) return(dr)
} }
#' @param nObs as describet in the Paper. #' @param nObs as described in the Paper.
#' @param X Data #' @param X Data
#' @param Y Responces #' @param Y Responces
#' @param nObs Like in the paper. #' @param nObs Like in the paper.

View File

@ -142,7 +142,7 @@ cve_rcg <- function(X, Y, k,
# Reset beta if needed. # Reset beta if needed.
if (loss.prime < 0) { if (loss.prime < 0) {
# Compute `beta` as describet in paper. # Compute `beta` as described in paper.
beta.FR <- (norm(A, type = 'F') / norm(A.last, type = 'F'))^2 beta.FR <- (norm(A, type = 'F') / norm(A.last, type = 'F'))^2
beta.PR <- sum(A * (A - A.last)) / norm(A.last, type = 'F')^2 beta.PR <- sum(A * (A - A.last)) / norm(A.last, type = 'F')^2
if (beta.PR < -beta.FR) { if (beta.PR < -beta.FR) {

View File

@ -1,7 +1,7 @@
#' Generates test datasets. #' Generates test datasets.
#' #'
#' Provides sample datasets. There are 5 different datasets named #' Provides sample datasets. There are 5 different datasets named
#' M1, M2, M3, M4 and M5 describet in the paper references below. #' M1, M2, M3, M4 and M5 described in the paper references below.
#' The general model is given by: #' The general model is given by:
#' \deqn{Y ~ g(B'X) + \epsilon} #' \deqn{Y ~ g(B'X) + \epsilon}
#' #'

View File

@ -30,7 +30,7 @@ See: \code{\link{formula}}.}
\item{Y}{Responces} \item{Y}{Responces}
\item{nObs}{as describet in the Paper.} \item{nObs}{as described in the Paper.}
\item{k}{guess for SDR dimension.} \item{k}{guess for SDR dimension.}

View File

@ -28,7 +28,7 @@ List with elements
} }
\description{ \description{
Provides sample datasets. There are 5 different datasets named Provides sample datasets. There are 5 different datasets named
M1, M2, M3, M4 and M5 describet in the paper references below. M1, M2, M3, M4 and M5 described in the paper references below.
The general model is given by: The general model is given by:
\deqn{Y ~ g(B'X) + \epsilon} \deqn{Y ~ g(B'X) + \epsilon}
} }