Professur für Biostatistik und Medizinische Biometrie
Sekretariat AG Biostatistik und Medizinische Biometrie
charlene.baumann@uni-bielefeld.de
Wissenschaftliche Mitarbeiterin
sandra.landwehr@uni-bielefeld.de
Wissenschaftlicher Mitarbeiter
juergen.rodenkirchen@uni-bielefeld.de
sebastian.buescher@uni-bielefeld.de
Discrete Choice Modells
Discrete choice models involve modelling the choice probabilities of a finite set of choices. Classical applications include transport mode choice, political choice, or product choice.
In particular, I deal with the composite marginal likelihood (CML) estimation of so-called probit models. In CML estimation, the product of the probabilities of pairs of observations is used instead of the joint probability of all observations for a person. Doing so reduces the necessary computational effort considerably. This comes at the expense of the statistical efficiency of the estimator.
My current focus is on
1. developing graphical diagnostic tools for the CML estimation,
2. testing for time dependencies within the model using the gradient contributions of the observation pairs,
3. reducing the loss of efficiency by using weighting strategies, and
4. making the estimator more robust to misspecification through the usage of weighting strategies.
Wissenschaftliche Mitarbeiterin
antoniya.dineva@uni-bielefeld.de
Wissenschaftliche Mitarbeiterin
kira.baginski@uni-bielefeld.de
Wissenschaftlicher Mitarbeiter
ferdinand.stoye@uni-bielefeld.de
Method development for the meta-analysis of diagnostic test accuracy studies