Applied Statistics

Possible Master's Thesis

Are you interested in writing a Master's thesis within the broad topic of Applied Statistics?

On this page we list some projects opportunities offered by our group. These are often rough outlines and will be further tailored to your needs and your background.

Contact us a few months early in case you are interested.

To improve your experience during the six months of thesis work, we set up quite a structured system guiding you through the thesis. Here are some of the boundaries:

  • The report has to be written in LaTeX with knitr.
  • All work has to be done in a reproducible framework. We adhere to these guidelines [link].
  • You present your work in an intermediate and in a final presentation.
  • We provide you with a starting repository in Git containing additional information about available resources, template for the report etc.

With these guidelines we are looking forward to effective and fruitful cooperation between you and our group.

Of course, if you have never worked with any of the above mentioned report generating systems,  you are no less welcome in our group and we will support you in acquiring the necessary skills to complete your thesis. We are also open to other engines for dynamic report generation (Jupiter/knitpy/...). 

List of topics

Clustering analysis of longitudinal dataset

The goal of this interdisciplinary Master's thesis is to perform a comprehensive literature review/screening for assessing the state of the art regarding the clustering analysis of a longitudinal dataset in observational animal study. Emphasis is put on using the R package kml. If possible we aim to perform a survival analysis with classical Cox models.

More details can be found here.

PIG DATA: transdisciplinary approach for health analytics of the Swiss swine industry

Bayesian network (BN) modeling is particularly useful to distinguish indirectly from directly associated variables, which are a promising target for intervention. Besides offering a nice display of the data, BN modeling can deliver structural quantitative information such as sets of variables that shields effect of index variable or variable importance for further modeling procedure. The goal of this Master thesis is to perform a comprehensive Bayesian network analysis on the PIG DATA using the R package abn.

More details can be found here.

Functional data ANOVA

The concept of an Analysis of Variance (ANOVA) for single observations is straightforward and quite intuitive from a statistical or geometrical point of view. However, if we observe an entire function instead, the concept needs to be extended.

Simulating nonstationary Gaussian random fields on the sphere

The goal of this project is to explore a novel turning band method to simulate nonstationary Gaussian random fields on the sphere and to compare it with existing approaches.
 

Automatic spatial plausibility testing of modeled height of new snow

The main objective is the development of an algorithm, which allows for automatic spatial plausibility
testing of modeled height of new snow. Furthermore, suitable statistical measures
should be proposed and evaluated, which might be used for plausibility testing. The final goal is a plausibility
flag for height values for each time step and weather station.

More information can be found here.

Cholesky factorization of sparse matrices

The main objective is a (numerical) complexity analysis of the Cholesky factorization implemented in the R package spam and compare the performence to alternative factorizations. Ideally, an existing approximate minimal degree algorithm will embedded in the spam environment.