Lectures

A link to lectures given at TUWIEN in 2009/10 is provided HERE.

List of Lectures:

Statistical Computing Introduction to statistical data analysis using modern computer methods with an emphasis on applied practical work. Solving data analysis problems and turning new ideas and methods into software.
Computational Statistics Exploration and analysis of statistical data. Descriptive statistics, comparison of groups of data, analysis of variance, regression.
Statistical Simulation and Computerintensive Methods Generation of Random Numbers, statistical simulation, high-performace computing and resampling methods using modern statistical software.
Cluster Analysis and Methods for Unsupervised Learning Theoretical concepts and application of cluster analysis and methods for unsupervised learning. Practical Applications using the software system
R.
Legal Aspects of Statistical Methods Realization of legal consequences when applying statistical methods. Applications in Official Statistics and Biostatistics.
Official Statistics Statistical methods in official statistics. Editing, Imputation, Complex Sampling Designs, Small Area Statistics, Statistical Disclosure Control
Seminars, Practicals Possible topics: Analysing fMRI data with R, Statistical Disclosure Control, Imputation, Synthetic Data Generation, etc.
Probability Theory and Statistics Practical courses
Lectures at Statistics Austria Lectures in Probability Theorie, Explorative Data Analysis, Visualization, Sampling Theory, Variance Estimation, Regression, Statistical Disclosure Control, Editing and Imputation.

For professional courses, see data-analysis OG .

Possible topics for Practicals, Seminars or Master Thesis:

  • Analysing brain slice data. A detailed description of the content can be found HERE (please, have a look at Figure 1 to get an impression of the clustering problem). The diploma thesis will be made in close cooperation with the Institute of Neurology, Medical University of Vienna. It is possible to continue to work by a PhD thesis.
  • Statistical Disclosure Control (SDC). SDC becomes more and more popular since data privacy must be ensured in many situations. The aim is to implement a global risk measurement framework via log-linear models in R. For this topic, the theory is well described, but an implementation in free and open-source software is missing. After reading few papers, the aim is to apply existing functions (e.g., the function loglin in R-package MASS) in R to solve this problem.
  • Imputation under Editing Constraints. Package VIM includes an algorithm called IRMI (iterative robust model-based imputation) which (multiple) imputes missing values in data robustly. However, often some combination of values should not occur in the data. The aim is to respect existing editing rules when imputing with IRMI.
  • Development and enhancement of graphical user interfaces using RGtk2, gWidgetsRGtk2 or tcl/tk.
  • C and/or C++. Transforming native R-code to C or C++.
  • Run simulations for the AMELI-project. The aim is to compare methods for estimation the laeken indicators (used for poverty measurement in Europe) using R-package simFrame.
  • At University of Natural Resources and Applied Life Sciences, Vienna (contact person at TU WIEN: Theresa Scharl ): Statistical Modeling of Bio-Processes. (P.S.: this diploma thesis is not supervised by any staff of our department.)
  • Comments are closed.