Eine Client/Server-Architektur für statistische Visualisierungen und Analyse von Finanzdaten

Statistical visualizations form an important instrument of the data analysis. In this thesis
a client/server-architecture for these visualizations is introduced. The architecture is
based on the components, whereby two of them contain a graphic-library for statistical
visualizations. The functions of each component are presented and their communication
is described. The prototype introduced in this work can be used on web pages,
in particular e-learning. The communication is made by the middleware CORBA. The
components were implemented in C++ and JAVA.
In the second part of the thesis models for analysis of finance data with student
t-distributions, especially a timeseries, are introduced. In the multivariate case models
depending on Copulas are presented. Copulas are defined, Estimators and simulationsalgorithms
are given. Particularly the question
”Which copula is the right one?“is discussed.