Information Technology Reference
In-Depth Information
Figure 5.13
Normal projection biplot axis with markers for
µ
=
2, 3 and 4 for original
variable
Y
.
In Figure 5.13 the markers
µ
=
2, 3 and 4 for original variable
Y
are shown on
the
L
∩
N
intersection spaces
l
(
2
)
z
=
m
(
2
)
,
l
(
3
)
z
=
m
(
3
)
and
l
(
4
)
z
=
m
(
4
)
.The
normal prediction biplot axis calculated according to (5.19) for a series of
µ
values is
shown in Figure 5.14.
The two nonlinear prediction biplot trajectories for normal projection are shown in
the biplot in Figure 5.15. Orthogonal projection from point P onto the axis
X
results in
a value of
5.6
and for
Y
we read off the value
2
.
In Section 5.6 we discuss the implementation of normal projection in our
R
function
Nonlinbipl
.
5.4.2.2 Circular projection
An alternative to orthogonal projection is to define the marker
on the intersection space
L
∩
N
as the orthogonal projection of the origin O (intersection of the scaffolding
axes) onto
L
∩
N
. This is illustrated in Figure 5.16. The equation of the intersection
space is
l
(µ)
z
=
t
(µ)
with intercept
t
(µ)/
l
2
(µ)
. To obtain the point P, the orthogonal
µ