vegan (version 1.17-12)

ordihull: Display Groups or Factor Levels in Ordination Diagrams

Description

Functions to add convex hulls, `spider' graphs, ellipses or cluster dendrogram to ordination diagrams. The ordination diagrams can be produced by vegan plot.cca, plot.decorana or ordiplot.

Usage

ordihull(ord, groups, display = "sites", draw = c("lines","polygon", "none"),
         show.groups, label = FALSE,  ...)
ordiellipse(ord, groups, display="sites", kind = c("sd","se"), conf,
         draw = c("lines","polygon", "none"), w = weights(ord, display),
         col = NULL, show.groups, label = FALSE, ...)
ordispider(ord, groups, display="sites", w = weights(ord, display),
         show.groups, label = FALSE, ...)
ordicluster(ord, cluster, prune = 0, display = "sites",
         w = weights(ord, display), ...)

Arguments

ord

An ordination object or an ordiplot object.

groups

Factor giving the groups for which the graphical item is drawn.

display

Item to displayed.

draw

Use either lines or polygon to draw the lines. Graphical parameters are passed to both. The main difference is that polygons may be filled

show.groups

Show only given groups. This can be a vector, or TRUE if you want to show items for which condition is TRUE. This argument makes it possible to use different colours and line types for groups. The default is to show a

label

Label the groups by their names in the centroid of the object. ordiellipse and ordihull use standard text, and others use

w

Weights used to find the average within group. Weights are used automatically for cca and decorana results, unless undone by the user. w=NULL<

kind

Whether standard deviations of points (sd) or standard deviations of their (weighted) averages (se) are used.

conf

Confidence limit for ellipses, e.g. 0.95. If given, the corresponding sd or se is multiplied with the corresponding value found from the Chi-squared distribution with 2df.

cluster

Result of hierarchic cluster analysis, such as hclust or agnes.

col

Colour of ellipses or ellipse fills in ordiellipse. For other functions the effect depends on the underlining functions this argument is passed to.

prune

Number of upper level hierarchies removed from the dendrogram. If prune $>0$, dendrogram will be disconnected.

...

Parameters passed to graphical functions or to scores to select axes and scaling etc.

Details

Function ordihull draws lines or polygons for the convex hulls found by function chull encircling the items in the groups. Function ordiellipse draws lines or polygons for dispersion ellipses using either standard deviation of point scores or standard error of the (weighted) average of scores, and the (weighted) correlation defines the direction of the principal axis of the ellipse. An ellipsoid hull can be drawn with function ellipsoidhull of package cluster.

Functions ordihull and ordiellipse return the invisible plotting structure. In ordihull this is a list of coordinates of the hulls (which can be extracted with scores), and in ordiellipse a list of covariance matrices and scales used in drawing the ellipses. These result objects have a summary method that returns the coordinates of the centres of the ellipses or hulls and their surface areas in user units. The centres of the hulls may differ from the location of the label which is the centre of the points instead of the centre of the polygon. With draw = "none" only the result object is returned and nothing is drawn.

Function ordispider draws a `spider' diagram where each point is connected to the group centroid with segments. Weighted centroids are used in the correspondence analysis methods cca and decorana or if the user gives the weights in the call. If ordispider is called with cca or rda result without groups argument, the function connects each `WA' scores to the corresponding `LC' score. If the argument is a (invisible) ordihull object, the functin will connect the points of the hull to their centroid.

Function ordicluster overlays a cluster dendrogram onto ordination. It needs the result from a hierarchic clustering such as hclust or agnes, or other with a similar structure. Function ordicluster connects cluster centroids to each other with line segments. Function uses centroids of all points in the clusters, and is therefore similar to average linkage methods.

See Also

The functions pass parameters to basic graphical functions, and you may wish to change the default values in lines, segments and polygon. You can pass parameters to scores as well. Underlying function for ordihull is chull.

Examples

data(dune)
data(dune.env)
mod <- cca(dune ~ Management, dune.env)
attach(dune.env)
## pass non-graphical arguments without warnings
plot(mod, type="n", scaling = 3)
## Catch the invisible result of ordihull...
pl <- ordihull(mod, Management, scaling = 3, label = TRUE)
## ... and find centres and areas of the hulls
summary(pl)
## use ordispider to label and mark the hull
plot(mod, type = "n")
pl <- ordihull(mod, Management, scaling = 3)
ordispider(pl, col="red", lty=3, label = TRUE )
## ordispider to connect WA and LC scores
plot(mod, dis=c("wa","lc"), type="p")
ordispider(mod)
## Other types of plots
plot(mod, type = "p", display="sites")
ordicluster(mod, hclust(vegdist(dune)), prune=3, col = "blue")
plot(mod, type="n", display = "sites")
text(mod, display="sites", labels = as.character(Management))
pl <- ordiellipse(mod, Management, kind="se", conf=0.95, lwd=2, col="blue")
summary(pl)