[Vegan-commits] r1666 - in pkg/vegan: inst man
noreply at r-forge.r-project.org
noreply at r-forge.r-project.org
Thu Jun 30 17:13:09 CEST 2011
Author: jarioksa
Date: 2011-06-30 17:13:09 +0200 (Thu, 30 Jun 2011)
New Revision: 1666
Modified:
pkg/vegan/inst/ChangeLog
pkg/vegan/man/monoMDS.Rd
Log:
help for monoMDS
Modified: pkg/vegan/inst/ChangeLog
===================================================================
--- pkg/vegan/inst/ChangeLog 2011-06-29 08:43:47 UTC (rev 1665)
+++ pkg/vegan/inst/ChangeLog 2011-06-30 15:13:09 UTC (rev 1666)
@@ -17,6 +17,8 @@
the update() in 1.18-31 -- problems were anticipated in ChangeLog,
and they were realized.
+ * monoMDS: text for a proper help (Rd file).
+
Version 1.18-33 (closed June 21, 2011)
* adonis: made faster (much faster in tests) by polishing the
Modified: pkg/vegan/man/monoMDS.Rd
===================================================================
--- pkg/vegan/man/monoMDS.Rd 2011-06-29 08:43:47 UTC (rev 1665)
+++ pkg/vegan/man/monoMDS.Rd 2011-06-30 15:13:09 UTC (rev 1666)
@@ -28,20 +28,21 @@
\item{model}{MDS model: \code{"global"} is normal non-metric MDS
with a monotone regression, \code{"local"} is non-metric MDS with
separate regressions for each point, \code{"linear"} uses linear
- regression, and \code{"hybrid"} uses linear MDS for
- dissimilarities below the threshold, and global non-metric MDS for
- dissimilarities above the threshold. }
+ regression, and \code{"hybrid"} uses linear regression for
+ dissimilarities below a threshold in addition to monotone
+ regression. See Details.}
- \item{threshold}{Dissimilarity above which monotone regression is
- used instead of linear regression. }
+ \item{threshold}{Dissimilarity below which linear regression is
+ used alternately with monotone regression. }
+
\item{maxit}{Maximum number of iterations.}
\item{weakties}{Use primary or weak tie treatment, where equal
- observed dissimilarities care allowed to have different fitted
+ observed dissimilarities are allowed to have different fitted
values. if \code{FALSE}, then secondary (strong) tie treatment are
used, and tied values are not broken.}
- \item{stress}{Use Kruskal's stress type 1 or 2.}
+ \item{stress}{Use stress type 1 or 2 (see Details).}
\item{scaling}{Scale final scores to unit root mean squares.}
@@ -60,33 +61,112 @@
\code{monoMDS}, passed to graphical functions in \code{plot}.).}
}
-\details{
- This is the first version of the function, and still under construction.
-}
+\details{ There are several versions of non-metric multidimensional
+ scaling in \R, but \code{monoMDS} offers the following unique
+ combination of features:
+ \itemize{
+
+ \item \dQuote{Weak} treatment of ties (Kruskal 1964a,b), where
+ tied dissimilarities can be broken in monontone regression. This
+ is especially important for cases where compared sites share no
+ species and dissimiliraties are tied to their maximum value of
+ one, and breaking these tied values is crucial on long gradients
+ to avoid curved artefacts. Functions
+ \code{\link[smacof]{smacofSym}} (\pkg{smacof} package) and
+ \code{\link[ecodist]{nmds}} (\pkg{ecodist} package) also have
+ adequate tie treatment.
-\value{
- A complicated object (under construction).
+ \item Handles missing values in a meaningful way.
+
+ \item Offers \dQuote{local} and \dQuote{hybrid} scaling in
+ addition to usual \dQuote{global} NMDS (see below).
+
+ \item Uses fast compiled code (\code{\link[MASS]{isoMDS}} of the
+ \pkg{MASS} package also uses compiled code).
+
+ }
+
+ Function \code{monoMDS} uses Kruskal's (1964b) original monotone
+ regression to minimize the stress. There are two alternatives of
+ stress: Kruskal's (1964a,b) original or \dQuote{stress 1} and an
+ alternative version or \dQuote{stress 2} (Sibson 1972). Both of
+ these stresses can be expressed with a general formula
+
+\deqn{s^2 = \frac{\sum (d - \hat d)^2}{\sum(d - d_0)^2}}{stress^2 = sum (d-dhat)^2/ sum (d-dnull)^2}
+
+ where \eqn{d} are the observed dissimilarities, \eqn{\hat d}{dhat}
+ are the fitted ordination distances, and \eqn{d_0}{dnull} are the
+ ordination distances under null model. For \dQuote{stress 1}
+ \eqn{d_0 = 0}{dnull = 0}, and for \dQuote{stress 2} \eqn{d_0 = \bar
+ d}{dnull = dbar} or mean dissimilarities.
+
+ Function \code{monoMDS} can fit several alternative NMDS variants
+ that can be selected with argument \code{model}. The default
+ \code{model = "global"} fits global NMDS, or Kruskal's (1964a,b)
+ original NMDS similar to \code{\link[MASS]{isoMDS}} (\pkg{MASS}),
+ \code{\link[smacof]{smacofSym}} (\pkg{smacof}) or
+ \code{\link[ecodist]{nmds}} (\pkg{ecodist}). Alternative
+ \code{model = "local"} fits local NMDS where independent monotone
+ regression is used for each point (Sibson 1972). Alternative
+ \code{model = "linear"} fits a linear MDS. This fits a linear
+ regression instead of monotone, and is not identical to metric
+ scaling or principal coordinates analysis that performs an
+ eigenvector decomposition of dissimilarities (Gower
+ 1966). Alternative \code{model = "hybrid"} implements hybrid MDS
+ thatuses monotone regression for all points and linear regression
+ for dissimilarities below or at a \code{threshold} in alternating
+ steps (Faith et al. 1987). Function \code{\link{stressplot}} can be
+ used to display the kind of regression in each \code{model}.
+
}
+\value{ Returns an object of class \code{"monoMDS"}. The final scores
+ are returned in item \code{points} (function \code{scores} extracts
+ these results), and the stress in item \code{stress}. In addition,
+ there is a large number of other items (but these may change without
+ notice in the future releases). }
+
\references{
+
+ Faith, D.P., Minchin, P.E and Belbin, L. 1987. Compositional
+ dissimilarity as a robust measure of ecological
+ distance. \emph{Vegetatio} 69, 57--68.
+
+ Gower, J.C. (1966). Some distance properties of latent root and
+ vector methods uses in multivariate analysis. \emph{Biometrika}
+ 53, 325--328.
+
Kruskal, J.B. 1964a. Multidimensional scaling by optimizing
goodness-of-fit to a nonmetric hypothesis. \emph{Psychometrika}
29, 1--28.
Kruskal, J.B. 1964b. Nonmetric multidimensional scaling: a numerical
method. \emph{Psychometrika} 29, 115--129.
+
+ Minchin, P.R. 1987. An evaluation of relative robustness of
+ techniques for ecological ordinations. \emph{Vegetatio} 69,
+ 89--107.
+
+ Sibson, R. 1972. Order invariant methods for data
+ analysis. \emph{Journal of the Royal Statistical Society B} 34,
+ 311--349.
+
}
\author{
Peter R. Michin (Fortran core) and Jari Oksanen (R interface).
}
-\note{
- Still under development.
+\note{ This is the default NMDS function used in
+ \code{\link{metaMDS}}. Function \code{\link{metaMDS}} adds support
+ functions so that NMDS can be run like recommended by Minchin
+ (1987).
}
-\seealso{ \code{\link[vegan]{metaMDS}}, \code{\link[MASS]{isoMDS}},
- \code{\link[smacof]{smacofSym}}, \code{\link[ecodist]{nmds}}. }
+\seealso{ \code{\link[vegan]{metaMDS}} for the \pkg{vegan} way of
+ running NMDS, and \code{\link[MASS]{isoMDS}},
+ \code{\link[smacof]{smacofSym}}, \code{\link[ecodist]{nmds}} for
+ some alternative implementations of NMDS. }
\examples{
data(dune)
More information about the Vegan-commits
mailing list