Graphics Reference
In-Depth Information
Figure
.
.
One-year probabilities of default for different rating grades (Füser,
)
A, or BB. Each company is placed into a specific class depending on how well it fits
thedescription. Forexample,theAAA gradeisassociated with extremely strong,AA
withverystrong,Awithstrong,BBBwithgood,BBwithmarginal,Bwithweak,CCC
with very weak and CC with extremely weak financial security characteristics, while
C signals imminent default, and D default.
A certain range of scores and PDs belongs to each rating class. he ranges were
computed on the basis of historical data. To derive a PD for a newly scored company,
its score f is compared with the historical values of the f values for each class. Based
on how similar the scores are, the company is assigned to one particular class. he
PD of this class becomes the PD of the company.
Company bond ratings play an important role in determining the cost of debt
refinancing,sincetheyreflecttheprobabilityofdefaulting onthedebt(Fig.
.
).Note
thatthedifferencesbetweentheclassesintermsofPDsarenotthesame.Forexample,
the PD increases by
.
% or
-fold between classes BBB and B, but only by
.
or
-fold between classes AAA und A. he colours used to code PDs must be selected
so that the classes appear to be equidistant, no matter what the absolute PD is. his
can be achieved by using an appropriate colour scheme and colour distance scaling.
he use of the HLS colour scheme in combination with a logarithmic colour scaling
will be demonstrated in Sect.
.
.
The SVM Approach
4.2
he SVM (Vapnik,
) is a classification and regression (which is not applied in
thiscase)techniquethatisbasedonmarginmaximisation(Fig.
.
)betweentwodata
classes.hemarginistheregionbetween thehyperplanesboundingeachclasswhere
noobservation can lieinalinear perfectly separable case.heclassifier function used
