This image shows Julius Voggesberger

Julius Voggesberger

M.Sc.

Researcher
IPVS
Applications of Parallel and Distributed Systems
[Photo: Julius Voggesberger]

Contact

Universitätsstraße 38
70569 Stuttgart
Germany
Room: 2.365

Office Hours

On appointment

Subject

Working Area: Machine Learning, Ensemble Learning, Classification

In my PhD topic " Deployment and data-driven fusion of heterogeneous predictive models" I deal with the creation of optimized ensembles, i.e. the combination of multiple classification models. For this purpose, the influence of data, model diversity, etc. on ensembles shall be investigated. Furthermore, an automated procedure shall be developed to create optimized ensembles based on the discovered connections.

Citation Metrics

Publications

2023

Julius Voggesberger. "Optimierung von Klassifikator-Ensembles mit AutoML". In: GvDB. 2023.

Julius Voggesberger, Peter Reimann und Bernhard Mitschang. “Towards the Automatic Creation of Optimized Classifier Ensembles”. In: Proc. of the 25th Int. Conference on Enterprise Information Systems (ICEIS) – Volume 1. 2023, S. 614–621.

Michael Behringer, Dennis Treder-Tschechlov, Julius Voggesberger, Pascal Hirmer und Bernhard Mitschang u. a. “SDRank: A Deep Learning Approach for Similarity Ranking of Data Sources to Support User-Centric Data Analysis”. In: Proc. of the 25th Int. Conference on Enterprise Information Systems (ICEIS) – Volume 1. 2023, S. 419–428.

Thesis

J. Voggesberger, "AutoML für Ensembles: Erhöhung der Klassifikationsperformanz durch Optimierung der Modelldiversität und der Entscheidungsfusion"
Masterthesis. 2022.

J. Voggesberger, " Evaluation von Zwischenergebnissen in Entscheidungsbäumen"
Bachelorthesis. 2019.

Summer Term 2023

  • Advanced Seminar "Data Management for End-to-End Machine Learning"
    Supervision

Winter Term 2022/2023

  • Seminar "Entwicklung von Data-Science-Anwendungen"
    Supervision

Student Projects in Progress

  • Master Thesis:
    Meta-Learning for Classifier Ensemble Optimization
  • Master Thesis:
    Development of a method for data-driven prediction of the outputs of a cold forging simulation
To the top of the page