Seminar zur Data Analysis


For participation in the seminar please apply with your university email address to our web-based system

We will admit up to 15  Master- and Diploma students.

You can only apply tho the seminar, not to specific topics. After you have been admitted to the seminar, we will inform you per Email how you can submit preferences and we will then allocate topics.

„Data Analysis“ (Prof. Gebhardt)

In this seminar we will expand on the methods for prediction that you learned in the course Data Analysis. In particular, the topics focus on the application of machine learning techniques. Your paper and discussion is expected to introduce a technique and then apply it to a data set (the same for all topics). 

Possible topics are:

  • Shrinkage Methods: LASSO, Ridge Regression
  • Principal Components Regression andPartial Least Squares
  • Trees and Boosting
  • Random Forests
  • Neural Nets

You may suggest topics of your own!

The Seminar will be based on

Friedman, J., Hastie, T., and Tibshirani, R. (2008) The elements of statistical learning, 2nd Edition Springer 

 James, G., Witten, D., Hastie, T., and Tibshirani, R. (2013) An introduction to statistical learning Springer

Please, be aware that you need a thorough knowledge of Micro-econometrics and in particular inference for prediction  to successfully participate in this seminar. Therefore, this seminar is intended for students who have taken my course "Data Analysis" in the Summer Semester. In addition, you will have to apply the techniques to a data set. You can use any software you would like, but for quite a few topics R will be the only choice.


Lecture "Data Analysis: Discription, Inference and Causality"


Master students that have taken my course "Data Analysis"

Ablauf und Termine

Participants have to write a paper and give a presentation. The seminar will take place in January. There will be a meeting at the end of this semester in which topics (and the associated literature) will be allocated and you will get all necessary information.