Hi Guys,
First, English is not my native language so sorry if the question is
too difficult to understand. I can rephrase it if necessary.
I have 32-bit Windows Xp SP3, i use R version 2.12.0 (2010-10-15), and
the question is about Redundancy canonical analysis plot problem in
3D using VEGAN, RGL, SCATTERPLOT3D and SFSMISC.
I noticed the following code to explain as clearly as possible the
problems encountered.
I am not able to add " species scores " from RDA results in 3D plot
like i can in 2D plot.
With the code i used , i get errors like ;
ERROR dans t$sites : $ operator is invalid for atomic vectors,
Erreur dans match.arg (display, items) : 'arg' should be one of
xyz.convert, points3d, plane3d, box3d, points, arrows
I saw that I can use other functions, but it is impossible to extract
the 3D scores ( and add species to 3d plot ), Other fonctions works
but, add elements in a 2D space.
How can I add the "species scores" (sites and arrows are added
correctly) to the 3D plot as with the 2D plot.
Thank you for your help,
Sincerely,
Thomas
University of Quebec at Trois-Rivieres , Canada
Environmental science MSC student.
rouss...@uqtr.ca, thomasbeaum...@hotmail.com
Here is the noticed code I used;
----- CODE ----------------
# Load library
library(vegan)
library(rgl)
library(scatterplot3d)
library(sfsmisc)
# Import data
ALLMAS <- read.table("C:/data.txt", header=TRUE, sep="",
na.strings="NA", dec=".", strip.white=TRUE
# Define var.
X <-ALLMAS[,c(10,14,15,16,18,19,22)] # env. (7) var.
Y <-ALLMAS[,c(9,3:6)] # y. (4) var
Z <-ALLMAS[,c(25)] # Z. (1) var covariate
ENVIESRDA <-rda(Y,X,Z) # Redundancy canonical analysis
pl <- ordiplot3d(ENVIESRDA, angle=15, type="n") # Get $Points (1:2) ,
get $arrows (1:2)
text(pl, "arrows", col="blue", pos=3, cex = 1) # Text env. var in plot
text(pl, "points", col="red", pos=1, cex = 0.5) # Point "sites"
(y.var) in plot
text(pl, "species", col="blue", cex=0.8) # Text " species " in plot ,
# I get the following error message:
# Erreur dans match.arg(display, items) :
# 'arg' should be one of xyz.convert, points3d, plane3d,
box3d, points, arrows
# I saw that I can use other functions, but it is impossible to
extract the 3D scores. Other features add elements in a 2D space.
# Then I try the following code, which works well in 2D when I
extract the summary of the results of the RDA.
t <-summary(ENVIESRDA,scal=2) # t <-summary(pl,scal=2) ????
# I get ;
# Length Class Mode
# xyz.convert 1 -none- function
# points3d 1 -none- function
# plane3d 1 -none- function
# box3d 1 -none- function
# points 48 -none- numeric
# arrows 14 -none- numeric
# Points = sites but where is " species scores " ?
X =t$sites[,1] # ERROR dans t$sites : $ operator is invalid for atomic vectors
Y =t$sites[,2]
t$sites
summary(ENVIESRDA, step=1000, perm.max=1000) # Summary
anova(ENVIESRDA, step=1000, perm.max=1000) # Anova
# Draw arrows to connect the "sites" with each other before / after study.
# But I can't get the new coordinates of 3D graphics using the
function summary.
stn2= p.arrows(X[1],Y[1], X[13],Y[13],fill="red",width=1)
stn3= p.arrows(X[2],Y[2], X[14],Y[14],fill="yellow",width=1)
stn4= p.arrows(X[3],Y[3], X[15],Y[15],fill="yellow",width=1)
stn5= p.arrows(X[4],Y[4], X[16],Y[16],fill="red",width=1)
stn6= p.arrows(X[5],Y[5], X[17],Y[17],fill="yellow",width=1)
stn7= p.arrows(X[6],Y[6], X[18],Y[18],fill="red",width=1)
stn9= p.arrows(X[7],Y[7], X[19],Y[19],fill="yellow",width=1)
stn10= p.arrows(X[8],Y[8], X[20],Y[20],fill="yellow",width=1)
stn11= p.arrows(X[9],Y[9], X[21],Y[21],fill="red",width=1)
stn12= p.arrows(X[10],Y[10], X[22],Y[22],fill="yellow",width=1)
stn13= p.arrows(X[11],Y[11], X[23],Y[23],fill="red",width=1)
stn18= p.arrows(X[12],Y[12], X[24],Y[24],fill="red",width=1)
stn19= p.arrows(X[13],Y[13], X[25],Y[25],fill="red",width=1)
# END
-------------------------------------------------
Courriel expédié via https://courriel.uqtr.ca
[[alternative HTML version deleted]]
______________________________________________
R-help@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.
