Information Technology Reference
In-Depth Information
SE (0.07)
IRD (0.06)
3
14
16
4
5
2
1
10
15
9
17
12
18
7
8
19
11
13
EDM (1.00)
6
20
Ideal point
EDSA (0.05)
CM (0.99)
MM (0.05)
ALCO (0.11)
Figure 3.32
PCA biplot of the unscaled risk management data with smooth trend line.
In Figure 3.32 the biplot provides the risk manager with a single visual representation
showing the performance of the seven financial instruments. While
EDM
improved
and stabilized over the period, other instruments such as
CM
exhibited up and down
behaviour with a sudden change on about
day15
.
3.8.2 Quality as a multidimensional process
A manufacturing company is monitoring 15 different variables in a production process.
In an effort to quantify the overall product quality, this particular manufacturing com-
pany devised a quality index value. Throughout the period of a calendar month the 15
variables were measured during the production process. At the end of the month, the
means and standard deviations of the 15 selected variables were transformed into a sin-
gle quality index value in the interval [0,100]. We shall treat this transformation process
as a company-specific black box but emphasize that the biplot procedure described below
is independent of what is inside this black box. Although the index values coming out of
the black box are useful for top-level management reporting, they give no indication of
what the causes of a poor index value could be. The monthly mean values for January
2000 to March 2001 are given in Table 3.19. The variable names indicate eight different
