Applied Statistics

Possible Topics for Statistical Practice

If you are interested in completing a fun project, linked to a "real" statistical problem, e.g., consulting cases, statistical software implementations, methodology developments etc. please contact us.

The project can be completed in the framework of the module 'STA390 Statistical Practice', delivering between 4 to 6 ECTS. The some projects of the list below may require specific background for a smooth start.

 

List of topics

Evolution of the "swiss" dataset

The goal of this project is to create a small R package containing the entire "swiss" data.

Non-parametric bootstrapping with `abn`

The goal of this project is to develop some resampling guidelines in additive Bayesian network models to avoid overfitting.

Parametric bootstrapping with `abn`

The goal of this project is to develop a routine that constructs BUGs files for parametric bootstrapping with abn automatically.

Evaluation of covariance function approximation

The goal of this project is to numerically quantify approximations of the covariance functions on prediction of large spatial fields.

Clustering of national athems

This project is consists of a reanalysis in R of https://medium.com/@lucasdesa/text-clustering-with-k-means-a039d84a941b

Improving forecasts with a Kalman filter

The goal of this project is to implement and test a filter and forecast routine for positional data. More precisely, based on GPS mobile phone positions, the goal is to impute gaps (tunnels, missing signals, ...) and to smooth data.