Supervised theses
(Selection of) Supervised Theses
In order to give interested students an impression of how potential thesis topics could look like, we try to provide a somewhat representative overview on past theses:
2022
- A Comparative Evaluation of the Utility of Linguistic Features for Part-of-Speech-Tagging (BT, 2022)
- Application of neural topic models to twitter data from German politicians (BT, 2022)
- Evaluating pre-trained language models on partially unlabeled multilingual economic corpora (MT, 2022)
- Domain transfer across country, time and modality in multiclass-classification of political texts (BT, 2022)
2021
- Modern approaches in mortality modeling considering data enrichment and the impact of COVID-19 (MT, 2021)
- Aspect-based Sentiment Analysis: A Theoretical and Practical Comparison of Different Approaches (MT, 2021)
- Sequence to Sequence Models: Knowledge Tracing with Deep Learning (MT, 2021)
- Re-Evaluating GermEval 2017: Document-Level and Aspect-Based Sentiment Analysis Using Pre-Trained Language Models (MT, 2021)
2020
- Analyse des Bundesligafußballs anhand von Spielstatistiken (BT, 2020)
- Benchmarking down-scaled versions of large pre-trained language models (MT, 2020)
- Flexibilisierung des modellbasierten Boosting durch Resampling und Regularisierung (BT, 2020)
- Fake News Detection. Evaluating Unsupervised Representation Learning for Detecting Stances of Fake News (MT, 2020)
2019
- Positionsbezogene Leistungsdatenanalyse von Fußballspielern (MT, 2019)
- Text Mining - Anwendung moderner Machine Learning Verfahren zur Analyse von Kundenrezensionen (BT, 2019)
- A Hyper-Parameter-Tuned, Confidence-Weighted Collaborative Filtering Model for Implicit Feedback (BT, 2019)
2018