Information Technology Reference
In-Depth Information
50
55
60
65
70
75
LGAN
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008
Figure 2.41
Scatterplot of two head dimensions variables data set superimposed upon a
bivariate image plot of the data. Figure constructed using our function
Scatterdensity
such that a continuous background image plot is obtained.
group. In the next section, we provide an example of this. Overlaying a scatterplot
with density estimates for several groups generally leads to a graph that cannot be
easily comprehended. Therefore, we choose to use density estimates in the form of
image plots as background for our biplots overlaid with
α
-bags and/or
κ
-ellipses. It
follows from the above discussion that there are various methods for graphically sum-
marizing the shape of a bivariate configuration. In this topic, we concentrate on
α
-bags,
κ
-ellipses and image plots and will show by example that these are invaluable devices
for providing bivariate descriptive summaries when used together with biplots of large
data sets.
2.10 Buying by mail order catalogue data set revisited
The biplot given in Figure 2.28 is a two-dimensional approximation of the 90-dimensional
sample points in the mail catalogue data set. The matrix consisting of the two-dimensional
coordinates for plotting the biplot can be used as input to the R function
PCAbipl
to
