Theses
Master theses at the chair focus on inference methods for time series and panel data models as well as on financial forecasts. We strongly recommend that you have previously taken a seminar at the chair!
If you are interested in writing a thesis in the area of econometrics and statistics, please send an e-mail to Prof. Demetrescu with the following information:
- Your name, your enrolment number and your field of study.
- Your interest and, if possible, a first idea for a topic
- Your desired start date
Always use your university email address for communication!
Topics of interest concern mainly econometric forecasting (both theory and applications), but related proposals are also welcome.
Selected Previous Topics:
- On Tests of Equal Predictive Accuracy for Panel Data
- Extending the Quantilogram and Quantile Stability Tests
- Vector Quantile Regression for Electricity Market Forecasting
- How Robust Are Time-Constant BVARs to Structural Breaks?
- Uncovering periods of directional predictability of stock returns
- Statistically Efficient Neural Network based Forecasts of Stock Return Volatility
- Restoring the monotonicity of the power function under time-varying alternatives: Parametric and non-parametric approaches
- An empirical multivariate analysis of selected DJIA realized volatilities
- Forecasting Value at Risk Using Selected Data Mining Techniques: The S&P500 Case
- An IVX Approach To Panel Unit Roots Testing