Project Details
Projekt
GRK 1032: Statistical Modelling
Subject Area
Economics
Mathematics
Mathematics
Term
from 2004 to 2013
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 374400
The project fosters research in "Statistical Modelling", an interface between theory and practice as well as in different fields of applications of statistics (for instance biometrics, econometrics). Work in the project is divided into four research areas: survey of data for statistical modelling, basics for statistical modelling, empirical modelling, algorithms in statistical modelling. Interaction among these fields is the main focus of the project.
Natural phenomena and processes are not exactly describable. Merely the appearance of the phenomena can be percepted, only schemes can be imagined. Furthermore, the observed phenomena are subject to random fluctuations whereas you normally do not know the nature of these variations. A central element of empirical sciences is to approximate those phenomena and processes by statistical models.
Aim of the project is to find and analyse models able to describe and to analyse the interesting problems. Reality should be reproduced as accurate as possible without losing the handling of the methods. Thus we pursue four basic ideas:
Developing models is motivated by practical problems. For adequat descriptions mathematical-statistical knowledge must be combined with professional expertise from application. The models to construct are dependent on the respective question (to illustrate an incident or an occurence, to forecast, as explanation).
Models from different fields are often from the same structure. Theoretical further development of these basic models in order to generalise them yields new fields of application. The analysis of constraining assumptions also provides us with a better usability. A model needs a certain structure of data. For checking the performance of the model and increasing the validity of the model it is essential to find suitable models for the data survey (for instance from experiments).
The computer-aided implementation of statistical models gets more and more important. Adequate algorithms have to be found in order to apply those statistical models for huge amounts of data, simulation studies provide an algorithmic verification of the models.
Natural phenomena and processes are not exactly describable. Merely the appearance of the phenomena can be percepted, only schemes can be imagined. Furthermore, the observed phenomena are subject to random fluctuations whereas you normally do not know the nature of these variations. A central element of empirical sciences is to approximate those phenomena and processes by statistical models.
Aim of the project is to find and analyse models able to describe and to analyse the interesting problems. Reality should be reproduced as accurate as possible without losing the handling of the methods. Thus we pursue four basic ideas:
Developing models is motivated by practical problems. For adequat descriptions mathematical-statistical knowledge must be combined with professional expertise from application. The models to construct are dependent on the respective question (to illustrate an incident or an occurence, to forecast, as explanation).
Models from different fields are often from the same structure. Theoretical further development of these basic models in order to generalise them yields new fields of application. The analysis of constraining assumptions also provides us with a better usability. A model needs a certain structure of data. For checking the performance of the model and increasing the validity of the model it is essential to find suitable models for the data survey (for instance from experiments).
The computer-aided implementation of statistical models gets more and more important. Adequate algorithms have to be found in order to apply those statistical models for huge amounts of data, simulation studies provide an algorithmic verification of the models.
DFG Programme
Research Training Groups
Applicant Institution
Technische Universität Dortmund
Spokesperson
Professor Dr. Joachim Kunert
Participating Researchers
Professor Dr. Roland Fried; Professorin Dr. Ursula Gather; Professor Dr. Joachim Hartung (†); Professorin Dr. Katja Ickstadt; Professor Dr. Guido Knapp; Professor Dr. Walter Krämer; Dr. Uwe Ligges; Professor Dr. Jörg Rahnenführer; Professor Dr. Götz Trenkler; Professor Dr. Claus Weihs; Professor Dr. Rafael Weißbach
