Information Technology Reference
In-Depth Information
0.1
-0.01
-0.01
0.15
-0.04
RAC
-0.02
AtMr
0.02
-0.1
-0.02
0.1
0.05
CmRb
Gaut
CrJk
-0.02
-0.005
-0.05
-0.01
0.05
0.02
InAs
KZN
DrgR
CmRb
0.02
DrgR
0.2
AtMr
-0.02
0.01
0
0
WCpe
InAs
0.1
CmAs
PubV
Mrd
BRs
0
CmAs
-0.1
-0.01
Arsn
Rape
BNRs
Mpml
-0.02
NWst
FrSt
-0.05
Limp
ECpe
0.005
NCpe
0.05
0.01
0.02
-0.04
-0.05
-0.02
0.01
0.1
AGBH
0.02
0.02
Figure 7.8
Two-dimensional CA biplot of the 2007/08 crime data set resulting from
the SVD
R
−
1
/
2
=
U
V
and plotting
U
and
V
with argument
lambda
= TRUE
. Determination of predictions for the Western Cape is illustrated.
(
X
−
E
)
C
−
1
/
2
with the origin suggest that any associated row points will be behaving averagely except
for the weights for the collinear column points. Thus, it can be seen that the Western
Cape diverges from the average in respect of
DrgR
,
InAs
and
CmAs
. Similarly,
Gaut
is
unusual for
CrJk
,
RAC
,
CmRb
and
AtMr
.
It is interesting to compare the actual values of the lambda parameter obtained when
specifying
ca.variant = "ConRatioA"
and
ca.variant = "ConRatioB"
, respec-
tively. When
"ConRatioA"
is specified a lambda value of 1.0648 is obtained. This
shows that the specifying argument
lambda = TRUE
should not lead to a biplot that
differs much from the biplot obtained with the default
lambda = FALSE
setting. On the
other hand, a lambda value of 2.3072 is obtained when
ca.variant = "ConRatioB"
is specified together with
lambda = TRUE
. The effect of this value is obvious when
Figure 7.11 is compared with Figure 7.12.