Graphical Data Representation in Bankruptcy Analysis



16


W. K. Hardle, R. A. Moro, and D. Schafer

PD and Cumulative PD

50          100          150

Company rank


200


Fig. 12. PD (blue line) and cumulative PD (green line) estimated with the SVM
for a subsample of 200 observation from the Bundesbank data. The variables were
included into the model that achieved the highest AR: K5, K29, K7, K33, K18, K21,
K24 and K9. The higher is the score, the higher is the rank of a company.

into an RGB colour and plotted by the XploRe as a small filled quadrangle.
The quadranges evenly cover the whole area giving a continuous PD repre-
sentation. The contour lines can be also added to the graph as illustrated by
Figure 18.

7 Conclusion

In this chapter we demonstrated the application of graphical tools for variable
selection, data visualisation and financial information representation and dis-
cussed such essential aspects of graphical analysis as colour coding. We believe
that graphical analysis will have an increased importance as non-parametric
models, such as SVM, are becoming more and more popular. On the other
hand graphical representation can facilitate the acceptance of non-parametric
models in various areas, e.g. finance, medicine, sound and image processing,
etc. This will contribute to the development of those areas since non-linear
non-parametric models better represent reality and provide higher forecasting
accuracy.



More intriguing information

1. A Duality Approach to Testing the Economic Behaviour of Dairy-Marketing Co-operatives: The Case of Ireland
2. Regional differentiation in the Russian federation: A cluster-based typification
3. Investment in Next Generation Networks and the Role of Regulation: A Real Option Approach
4. Internationalization of Universities as Internationalization of Bildung
5. How do investors' expectations drive asset prices?
6. The name is absent
7. A production model and maintenance planning model for the process industry
8. Three Strikes and You.re Out: Reply to Cooper and Willis
9. The name is absent
10. The name is absent