Canonical variate analysis biplots - Understanding Biplots - page 152

Ta b l e 4 . 3 Wood anatomical measurements
obtained for a specimen of unknown origin.

Wood anatomical feature Mean values ( µ m)

VesD 134

VesL 375

FibL 1170

We now return to the problem of identifying the kind of wood used in the manufac-
ture of the controversial chair. Using mainly qualitative, microscopic wood anatomical
features, the sample was initially identified as belonging to the family Lauraceae and
then as a representative of the genus Ocotea . In the second step of the classification, a
small inconspicuous sample of wood from the chair was obtained and the microscopic
measurement values of FibL , Ve s L and Ve sD were used to calculate a mean value for
each variable similar to the methods used for collecting the data set in Table 3.9. These
values are given in Table 4.3.
In Figure 4.4 classification regions are added to the CVA biplot of Figure 4.3. Each
region contains the points nearest the relevant mean. The interpolated position of the

VesD

180

300

160

Unknown specimen

140

350

1000

120

1200

1400

1600

1800

400

FibL

100

80

Obul; n = 20
Oken; n

450

= 7. (C hull)

Opor; n = 10

VesL

Figure 4.4 CVA biplot using variables FibL , Ve s L and Ve sD of the Ocotea data set.
Classification regions have been added. Group means are indicated by solid symbols in
the legend. The black star is the interpolated position of the specimen of unknown origin.
0.95-bags are shown.