An embedded and charset-unspecified text was scrubbed... Name: not available URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20130227/ab8bded7/attachment.pl>
best ordination method for binary variables
2 messages · Marco, David L Carlson
1 day later
It is always useful to look at the data in multiple ways. The unique() function will remove the duplicates in your data so that isoMDS will work:
set.seed(42) x <- matrix(sample(0:1, 20, replace=TRUE), 10, 2) x
[,1] [,2] [1,] 1 0 [2,] 1 1 [3,] 0 1 [4,] 1 0 [5,] 1 0 [6,] 1 1 [7,] 1 1 [8,] 0 0 [9,] 1 0 [10,] 1 1
unique(x)
[,1] [,2] [1,] 1 0 [2,] 1 1 [3,] 0 1 [4,] 0 0 ---------------------------------------------- David L Carlson Associate Professor of Anthropology Texas A&M University College Station, TX 77843-4352
-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
project.org] On Behalf Of marco milella
Sent: Wednesday, February 27, 2013 6:09 AM
To: r-help at r-project.org
Subject: [R] best ordination method for binary variables
Dear all,
I'm analyzing a dataset (A) of 400 cases with 11 binary variables.
Unfortunately, several (actually a lot) of cases are identical. NA are
also
present.
I want to to plot distances between cases.
For this, I obtained a distance matrix by dist(A, method="binary"). I
then
analyzed the obtained distance via Principal coordinate analysis with
cmdscale(). Results are fine.
However, do you think this is a wrong approach? After reading the
literature and previous posts, I noticed that non metrical MDS (via
isoMDS
or metaMDS) could be a more correct choice.
The problem is that, when trying this methods, I immediately get
problems
due to the identity between several of mycases or the presence of NA.
Typical error messages are
*"Error in isoMDS(DistB, k = 3) : zero or negative distance between
objects
1 and 2"*
or
*"Error in if (any(autotransform, noshare > 0, wascores) && any(comm <
0))
{ : missing value where TRUE/FALSE needed*
*In addition: Warning message:*
*In Ops.factor(left, right) : < not meaningful for factor"*
Do you think Principal coordinate analysis on a binary distance matrix
is a
decent strategy?
Thanks for any suggestion
marco
[[alternative HTML version deleted]]
______________________________________________ R-help at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting- guide.html and provide commented, minimal, self-contained, reproducible code.