Mathematical modeling is a fundamental tool of science. Models may be applied to extract knowledge from data, to combine available information, or to make predictions about the future states of a system. In any case models need to be tailored to the scientific questions at hand. Ideally a model utilize all available information. This includes (often expensive field or lab) data but also system understanding and expert opinions.
As a statistician my interests lay in constructing adequate models by means of statistics, machine learning and applied mathematics to investigate scientific hypotheses and support decision makers.
Besides the mathematical and technical aspects of modeling I’m also very interested in communication, which is an important element of successful collaboration. To be able to apply a model successful, users must understand the underlying model assumptions, how to interpret the results, and know the limits of a model.
Obtaining scientific data from field observations or experiments is a laborious and expensive task – my aim is to facilitate the best possible use of this hard-earned data.
Methods and Tools
Some topics and methods I work with or I am interested in:
- Bayesian Inference
- Gaussian Processes
- Data assimilation(Deep)
- Machine Learning
- Artificial Neuronal Networks
- Uncertainty Quantification
- Causal Inference
- Graphical (hierarchical) Models
For implementation I use among others Julia, R, Python, STAN, Emacs.