In applied disciplines contemporary data sets are often of increased size, detail and complexity. From a statistical point of view this is both a blessing and a curse imposing additional mathematical, computational and applied challenges to obtain realistic and reliable statistical models. In this project we want to combine the concepts of Bayesian regularization, machine learning and semi-and nonparametric methods to build efficient novel models that are extremely useful in the context of calibrated density regression. In contrast to classical statistical methods, Bayesian statistics offers a more direct expression of uncertainty through a prior distribution and allows updating knowledge.
During my assistant professorship, I have the goals to establish as an independent researcher with an international network of collaborations, to clearly improve my teaching skills and to grow into a successful group head, supervisor and mentor of the younger generation consisting of students, PhDs and postdoctoral researchers. The KT Boost Fund makes important contributions trough mentoring programmes and because it allows me to employ qualified scientific staff for reaching the project goals. I furthermore have the possibiliy to visit important conferences and collaboration partners in my field.
... would learn Spanish properly, read more novels and travel more often to Australia.
Name: Prof. Dr. Nadja Klein
Research field: Bayesian Computational Methods, Bayesian Deep Learning, Machine Learning, Smoothing, Regularization and Shrinkage, Distributional Regression, Network Analysis, Spatial Statistics
Institution: Humboldt University of Berlin, Assistant Professor of Applied Statistics at the Chair of Statistics