Given a two-dimensional matrix of vegetation data the function derives a contingency table of counts (scores presenc-absence transformed) based on input classification of rows (the vegetation releves) and columns (the species). The cells of the contingency table are then adjusted to equal weight, followed by correspondence analysis (cca). Concentration of counts is measured and an ordination plotted.
Usage
aocc(veg, o.rgr, o.sgr,...)
aoc(veg, o.rgr, o.sgr)
## Default S3 method:
aocc(veg, o.rgr, o.sgr,...)
## S3 method for class 'aocc'
plot(x,...)
Arguments
veg
A data frame of vegetation releves (rows) by species (columns)
o.rgr
Group membership of rows given upon input
o.sgr
Group membership of columns given upon input
x
An object of class "aocc"
...
Further variables used for plotting
Details
These input parameters are typically generated by functions clust() and
cutree() in the cluster package. See example below.
Value
An output list of class "aocc" with at least the following items:
rgrscores
Ordination scores of releve groups
sgrscores
Ordination scores of species groups
eigvar
Eigenvalues of correspondence analysis
grand.total
Grand total of contingency table
MSCC
Mean square contingency coefficient, a measure of concentration
new.relorder
Order of rows after ordering groups according to 1. axis
new.sporder
Order of columns after ordering groups according to 1. axis
cont.table
The contingency table
Note
The analysis of lattice structure, described in some of the references, is not included in this function.
Author(s)
Otto Wildi
References
Feoli, E. & Orloci, L. 1979. Analysis of
concentration and detection of underying factors in structured tables.
Vegetatio 40: 49-54.
Orloci, L. & Kenkel, N. 1985. Introduction to
Data Analysis. International Co-operative Publ. House, Burtonsville,
MD.
Wildi, O. 2013. Data Analysis in Vegetation Ecology. 2nd ed. Wiley-Blackwell, Chichester.
Examples
# First, groups of releves are formed
require(vegan)
dr<- vegdist(nveg^0.5,method="bray") # dr is distance matrix of rows
o.clr<- hclust(dr,method="ward") # this is clustering
o.rgr<- cutree(o.clr,k=3) # 3 row groups formed
# Now I group the columns of nveg (the species)
# the same way as for rows
ds<- vegdist(t(nveg^0.25),method="euclid")
o.cls<- hclust(ds,method="ward")
o.sgr<- cutree(o.cls,k=4) # 4 column groups formed
o.aocc<- aocc(nveg,o.rgr,o.sgr)
plot(o.aocc) # double scatter plot
# 3 row-, 4 column goups as points.
# If cluster analysis is not used but classification is input by row and
# column to be processed by aocc():
o.rgr<- c(1,2,1,3,2,3,1,2,3,1,3)
o.sgr<- c(1,1,2,2,1,3,4,3,1,1,1,3,3,1,1,4,4,4,4,1,3)
o.aocc<- aocc(nveg,o.rgr,o.sgr)
plot(o.aocc)